Omnidirectional vehicle

[0001] The present invention relates to an omnidirectional vehicle capable of moving on a floor surface.

[0002] There is proposed an inverted pendulum type vehicle, in which a base body assembled with a travel motion unit which travels on a floor surface and an actuator which drives the travel motion unit, is also assembled with a mounting unit of an object of transportation which is capable of freely tilting with respect to the perpendicular direction. For example, there is proposed a vehicle which is capable of functioning as the inverted pendulum type vehicle by the present applicant (refer to Published PCT International Applications WO/2008/132778 and Published PCT International Applications WO/2008/132779). In the vehicle, it is necessary to move the travel motion unit in a manner to move a fulcrum of the inverted pendulum, so as to maintain a tilt angle of the payload supporting part to a desired tilt angle (so as to prevent the payload supporting part from fall due to tilting).

[0003] Therefore, a control technology of the inverted pendulum type vehicle is proposed by the present applicant (refer to Japanese Patent No. 3070015).

[0004] To be more specific, the inverted pendulum type vehicle in which a base body of a vehicle assembled with a payload supporting part of a pedestrian as the mounting unit of the object of transportation, is provided so as to be freely tiltable about two axes, about an axis in a fore-and-aft direction and about an axis in a lateral direction, with respect to the ball-shaped travel motion unit, is the object of control. Further, a driving torque of a motor is sequentially determined so as to make a difference between a measured value of a tilt angle of the base body (=tilt angle of the payload supporting part) and a desired tilt angle closer to 0, and also to make the difference between a measured value of a velocity of the motor as an actuator (and consequently a moving velocity of the traveling motion unit) and a desired velocity closer to 0. Thereafter, the travel motion of the traveling motion unit is controlled via the motor, in accordance with the determined drive torque.

[0005] However, in the case where the operation of the motor is controlled without taking into consideration a posture of the floor surface, more specifically an inclination angle of the floor surface with respect to the horizontal surface, there is a possibility that a reaction force the traveling motion unit receives from the floor surface becomes excessively weak from the viewpoint of maintaining the behavior of the vehicle stable.

[0006] Therefore, the present invention aims to provide a vehicle capable of operating stably in a manner according to a posture of the floor surface.

[0007] To fulfill the object, the present invention provides an omnidirectional vehicle having a traveling motion unit capable of moving in all directions including a first direction and a second direction which are orthogonal to each other on a floor surface, an actuator which drives the traveling motion unit, a base body on which the traveling motion unit and the actuator are mounted, a two-dimensional floor inclination estimation component which is configured to estimate inclination angles of the floor surface in respective two horizontal directions orthogonal to each other as a two-dimensional floor inclination, and a control component which is configured to control the actuator, wherein the two-dimensional floor inclination estimation component comprises: an instantaneous provisional two-dimensional floor inclination calculation component which is configured to calculate an instantaneous provisional value of the two-dimensional floor inclination, on the basis of the motion state of the vehicle and either of a driving force of the actuator or a propelling force of the travel motion unit, and according to an equation of motion of the vehicle; a vehicle orientation estimation component which is configured to estimate an orientation of the vehicle or a rate of change thereof; and a filter component which is configured to calculate a new estimated value of the two-dimensional floor inclination, by coordinate converting at least one of a past estimated value of the two-dimensional floor inclination and the instantaneous provisional value of the two-dimensional floor inclination calculated by the instantaneous provisional two-dimensional floor inclination calculation component, using the orientation of the vehicle or the rate of change thereof estimated by the vehicle orientation estimation component, so as to conform the coordinate system of the past estimated value and the instantaneous provisional value, and thereafter calculating a weighed mean value or a low-pass filter value using the past estimated value and the instantaneous provisional value (a first aspect of the invention).

[0008] According to the vehicle of the present invention, the coordinate systems of the past estimate value and the instantaneous provisional value of the two-dimensional floor inclination are made to conform, by coordinate converting at least one of the “past estimate value” and the “instantaneous provisional value” of the two-dimensional floor inclination, using the orientation of the vehicle or the rate of change thereof. The “orientation” of the vehicle means at least either one azimuth angle of the first direction and the second direction in the world coordinate system. Thereafter, by calculating the weighed mean value or the low-pass filter value using the past estimated value and the instantaneous provisional value of the two-dimensional floor inclination, the new estimated value of the two-dimensional floor inclination is calculated.

[0009] As a result, even when at least either one of the driving force of the actuator or the propelling force of the traveling motion unit and the motional states of the vehicle changes to an extent that the instantaneous provisional value of the two-dimensional floor inclination changes abruptly, it becomes possible to avoid the situation where the changing amount of the new estimated value with respect to the past estimated value of the two-dimensional floor inclination becomes excessive. Therefore, the operation of the actuator may be controlled appropriately from the viewpoint of stabilizing the posture of the vehicle according to the inclination state of the floor surface, on the basis of the new estimated value of the two-dimensional floor inclination.

[0010] The phrase that the control device and the constituent components thereof are “configured to” execute a given arithmetic processing means that a CPU (arithmetic processing device) reads out data and program from a memory, and executes the arithmetic processing according to the program on the basis of the read-out data. The phrase “configured to” is a concept including that the same is “programmed to” do something.

[0011] In the omnidirectional vehicle of the first aspect of the invention, the control component may be configured to correct at least the movement in the second direction on the basis of the estimated value of the two-dimensional floor inclination calculated by the two-dimensional floor inclination estimation component (second aspect of the invention).

[0012] According to the vehicle of the above-explained configuration, the operation of the actuator is controlled so that the moving mode of the vehicle is appropriately adjusted from the viewpoint of stabilizing the posture of the vehicle at least in the second direction. For example, the direction with higher necessity of stabilizing the posture of the vehicle in view of the structure of the vehicle than in the “first direction” may be adapted as the “second direction”.

[0013] In the vehicle of the second aspect of the invention, the omnidirectional vehicle is a vehicle capable of tilting at least in the second direction, wherein the control component may be provided with a tilt detecting component which is configured to detect the tilting of the base body, and a tilt control component which is configured to stabilize the posture of the base body on the basis of the detected tilting of the base body and to correct the tilting of the base body at least in the second direction towards an elevated side of the floor, on the basis of the estimated value of the two-dimensional floor inclination calculated by the two-dimensional floor inclination estimation component (third aspect of the invention).

[0014] According to the vehicle of the above-explained configuration, in the case where the floor surface is a slant surface in the second direction, the tilting posture of the base body taking the vertical line as a reference is corrected to the uphill side of the slope (the higher surface side). By doing so, the tendency of the base body tilting to the downhill side of the slant surface by its own weight is alleviated or relieved, so that stabilization of the posture of the base body at least in the second direction, and consequently the stabilization of the posture of the vehicle is achieved.

[0015] In the omnidirectional vehicle of the third aspect of the invention, the control component may be configured to control the operation of the actuator so as to displace the center-of-gravity of the vehicle in an amount according to the tilt angle of the base body in the direction in which the base body is tilting, and to correct the tilt angle of the base body so as to generate the center-of-gravity movement of the vehicle equivalent to the displace amount of the ground contact surface of the travel motion unit generated by the inclination of the floor surface (fourth aspect of the invention).

[0016] According to the vehicle with the above-explained configuration, in the case where the floor surface is a sloped surface, the tilt angle of the base body is corrected to the uphill side of the slant surface as is mentioned above, and as a result the motion of the actuator is controlled so as to make the center-of-gravity of the vehicle move towards the uphill side of the slant surface. By doing so, the displacement amount of the ground contact surface towards the downhill side of the slope surface by the weight of the vehicle is cancelled, so that the stabilization of the position of the vehicle in the direction along the slope surface is achieved, and consequently the stabilization of the posture of the vehicle is achieved.

[0038] The following will describe a first embodiment of the present invention.

[0039] (Basic Configuration of the Vehicle)

[0040] First, referring to FIG. 1 to FIG. 6, the structure of an omnidirectional vehicle in the present embodiment will be described.

[0041] As illustrated in FIG. 1 and FIG. 2, an omnidirectional vehicle 1 in the present embodiment includes a payload supporting part 3 for an occupant (driver), a traveling motion unit (drive unit) 5 capable of traveling in all directions (two-dimensional all directions, including a fore-and-aft direction and a lateral direction) on a floor surface while being in contact with a floor surface, an actuator 7 which imparts, to the traveling motion unit 5, a motive power for driving the traveling motion unit 5, and a base body 9 on which the payload supporting part 3, the traveling motion unit 5, and the actuator 7 are mounted.

[0042] Here, in the description of the present embodiment, the “fore-and-aft direction” and the “lateral direction” mean the directions that coincide or substantially coincide with the fore-and-aft direction and the lateral direction, respectively, of the upper body of an occupant aboard the payload supporting part 3 in a normal posture. Incidentally, “the normal posture” is a posture envisaged in the design related to the payload supporting part 3, and it is a posture in which the trunk axis of the upper body of the occupant is oriented approximately in the vertical direction and the upper body is not twisted.

[0043] In this case, in FIG. 1, “the fore-and-aft direction” and “the lateral direction” are the direction perpendicular to the paper surface and the lateral direction of the paper surface, respectively. In FIG. 2, “the fore-and-aft direction” and “the lateral direction” are the lateral direction of the paper surface and the direction perpendicular to the paper surface, respectively. Further, in the description of the present embodiment, the suffixes “R” and “L” attached to reference numerals will be used to mean the correspondence to the right side and left side, respectively, of the vehicle 1.

[0044] The base body 9 is provided with a lower frame 11, to which the traveling motion unit 5 and the actuator 7 are installed, and a support frame 13 extendedly provided upward from the upper end of the lower frame 11.

[0045] A seat frame 15 extending toward the front from the support frame 13 is fixed to the top of the support frame 13. Further, the seat 3 on which an occupant sits is installed on the seat frame 15. In the present embodiment, the seat 3 serves as the payload supporting part for an occupant. Hence, the omnidirectional vehicle 1 in the present embodiment (hereinafter referred to simply as the vehicle 1) travels on a floor surface with an occupant seated on the seat 3.

[0046] Further, grips 17R and 17L to be grasped as necessary by the occupant seated on the seat 3 are disposed on the right and left of the seat 3. These grips 17R and 17L are secured to the distal portions of brackets 19R and 19L, respectively, which are provided extendedly from the support frame 13 (or the seat frame 15).

[0047] The lower frame 11 is provided with a pair of cover members 21R and 21L disposed to face each other in a forked shape with a gap therebetween in the lateral direction. The upper end portions (the forked portions) of these cover members 21R and 21L are connected through a hinge shaft 23 having a longitudinal axial center, so that one of the cover members 21R and 21L is relatively swingable about the hinge shaft 23 with respect to the other. In this case, the cover members 21R and 21L are biased by springs, which are not shown, in the direction in which the bottom end portions (the distal ends of the forked portions) of the cover members 21R and 21L narrow.

[0048] Further, a step 25R on which the occupant seated on the seat 3 rests his/her right foot and a step 25L on which the occupant rests his/her left foot are provided on the outer surfaces of the cover members 21R and 21L such that the steps extend out rightward and leftward, respectively.

[0049] The traveling motion unit 5 and the actuator 7 are disposed between the cover members 21R and 21L of the lower frame 11. The structures of the traveling motion unit 5 and the actuator 7 will be described with reference to FIG. 3 to FIG. 6.

[0050] The traveling motion unit 5 and the actuator 7 illustrated in the present embodiment have the same structures as those disclosed in, for example, FIG. 1 of patent document 2 mentioned above. Hence, in the description of the present embodiment, the aspects of the structures of the traveling motion unit 5 and the actuator 7 which are described in the aforesaid patent document 2 will be only briefly described.

[0051] In the present embodiment, the traveling motion unit 5 is a wheel assembly made of a rubber elastic material formed into an annular shape and has a substantially circular cross-sectional shape. This traveling motion unit 5 (hereinafter referred to as the wheel assembly 5) elastically deforms to be capable of rotating about a center C1 of the circular cross-section (more specifically, the circumferential line which passes the center C1 of the circular cross-section and which is concentric with the axial center of the wheel assembly 5), as indicated by an arrow Y1 in FIG. 5 and FIG. 6.

[0052] The wheel assembly 5 is disposed between the cover members 21R and 21L with an axial center C2 thereof (an axial center C2 orthogonal to the diametrical direction of the whole wheel assembly 5) oriented in the lateral direction, and comes in contact with a floor surface at the bottom end portion of the outer circumferential surface of the wheel assembly 5.

[0053] The wheel assembly 5 is capable of performing a motion of rotating about the axial center C2 of the wheel assembly 5 as indicated by an arrow Y2 in FIG. 5 (a motion of circumrotating on a floor surface) and a motion of rotating about the center C1 of the cross-section of the wheel assembly 5 by being driven by the actuator 7 (to be discussed in detail later). As a result, the wheel assembly 5 is capable of traveling in all directions on a floor surface by the motions combining the aforesaid rotating motions.

[0054] The actuator 7 is provided with a rotating member 27R and free rollers 29R interposed between the wheel assembly 5 and the right cover member 21R, a rotating member 27L and free rollers 29L interposed between the wheel assembly 5 and the left cover member 21L, an electric motor 31R serving as an actuator disposed above the rotating member 27R and the free rollers 29R, and an electric motor 31L serving as an actuator disposed above the rotating member 27L and the free rollers 29L.

[0055] The housings of the electric motors 31R and 31L are installed to the cover members 21R and 21L, respectively. Although not shown, the electric sources (batteries or capacitors) of the electric motors 31R and 31L are mounted on an appropriate place of the base body 9, such as the support frame 13 or the like.

[0056] The rotating member 27R is rotatively supported by the cover member 21R through the intermediary of a support axis 33R having a lateral axial center. Similarly, the rotating member 27L is rotatively supported by the cover member 21L through the intermediary of a support axis 33L having a lateral axial center. In this case, the rotational axial center of the rotating member 27R (the axial center of the support axis 33R) and the rotational axial center of the rotating member 27L (the axial center of the support axis 33L) are concentric with each other.

[0057] The rotating members 27R and 27L are connected to the output shafts of the electric motors 31R and 31L, respectively, through the intermediary of power transmission mechanisms, including functions as reducers, and rotatively driven by the motive power (torque) transmitted from the electric motors 31R and 31L, respectively. The power transmission mechanisms are, for example, pulley and belt system. More specifically, as illustrated in FIG. 3, the rotating member 27R is connected to the output shaft of the electric motor 31R through the intermediary of a pulley 35R and a belt 37R. Similarly, the rotating member 27L is connected to the output shaft of the electric motor 31L through the intermediary of a pulley 35L and a belt 37L.

[0058] Incidentally, the aforesaid power transmission mechanism may be constructed of, for example, a sprocket and a link chain, or may be constructed of a plurality of gears. As another alternative, for example, the electric motors 31R and 31L may be constructed such that the output shafts thereof are arranged to oppose the rotating members 27R and 27L so as to arrange the output shafts to be concentric with the rotating members 27R and 27L, and the output shafts of the electric motors 31R and 31L may be connected to the rotating members 27R and 27L, respectively, through the intermediary of reducers (e.g., planetary gear devices).

[0059] The rotating members 27R and 27L are formed in the same shapes as circular truncated cones, the diameters of which reduce toward the wheel assembly 5, and the outer peripheral surfaces thereof form tapered outer peripheral surfaces 39R and 39L.

[0060] A plurality of the free rollers 29R are arrayed about the tapered outer peripheral surface 39R of the rotating member 27R such that the free rollers 29R are arranged at regular intervals on the circumference concentric with the rotating member 27R. Further, these free rollers 29R are installed to the tapered outer peripheral surface 39R through the intermediary of the brackets 41R and rotatively supported by the brackets 41R.

[0061] Similarly, a plurality of free rollers 29L (of the same quantity as that of the free rollers 29R) are arrayed about the tapered outer peripheral surface 39L of the rotary member 27L such that the free rollers 29L are arrayed at regular intervals on the circumference concentric with the rotating member 27L. Further, these free rollers 29L are installed to the tapered outer peripheral surface 39L through the intermediary of the brackets 41L and rotatively supported by the brackets 41L.

[0062] The wheel assembly 5 is disposed concentrically with the rotating members 27R and 27L, and held between the free rollers 29R adjacent to the rotating member 27R and the free rollers 29L adjacent to the rotating member 27L.

[0063] In this case, as illustrated in FIG. 1 and FIG. 6, the free rollers 29R and 29L are disposed in postures in which the axial centers C3 thereof are inclined against the axial center C2 of the wheel assembly 5 and also inclined against the diametrical direction of the wheel assembly 5 (the radial direction connecting the axial center C2 and the free rollers 29R and 29L when the wheel assembly 5 is observed in the direction of the axial center C2 thereof). Further, in the aforesaid postures, the outer peripheral surfaces of the free rollers 29R and 29L, respectively, are pressed into contact aslant with the inner peripheral surface of the wheel assembly 5.

[0064] More generally speaking, the right free rollers 29R are pressed into contact with the inner peripheral surface of the wheel assembly 5 in postures in which a frictional force component in the direction about the axial center C2 (a frictional force component in the tangential direction of the inner periphery of the wheel assembly 5) and a frictional force component in the direction about the center C1 of the cross-section of the wheel assembly 5 (a frictional force component in the tangential direction of the circular cross section) can be applied to the wheel assembly 5 at a surface in contact with the wheel assembly 5 when the rotating member 27R is rotatively driven about the axial center C2. The same applies to the left free rollers 29L.

[0065] In this case, as described above, the cover members 21R and 21L are biased by the springs, which are not shown, in the direction for narrowing the bottom end portions (the distal ends of the forked portions) of the cover members 21R and 21L. Thus, the urging force holds the wheel assembly 5 between the right free rollers 29R and the left free rollers 29L, and the free rollers 29R and 29L are maintained in the press contact with the wheel assembly 5 (more specifically, the press contact state that enables a frictional force to act between the free rollers 29R and 29L and the wheel assembly 5).

[0066] (Basic Operation of the Vehicle)

[0067] In the vehicle 1 having the structure described above, when the rotating members 27R and 27L are rotatively driven at the same velocity in the same direction by the electric motors 31R and 31L, respectively, the wheel assembly 5 will rotate about the axial center C2 in the same direction as those of the rotating members 27R and 27L. This causes the wheel assembly 5 to circumrotate on a floor surface in the fore-and-aft direction and the whole vehicle 1 will travel in the fore-and-aft direction. In this case, the wheel assembly 5 does not rotate about the center C1 of the cross-section thereof.

[0068] Further, if, for example, the rotating members 27R and 27L are rotatively driven in opposite directions from each other at velocities of the same magnitude, then the wheel assembly 5 will rotate about the center C1 of the cross section thereof. This causes the wheel assembly 5 to travel in the direction of the axial center C2 thereof (i.e., in the lateral direction), thus causing the whole vehicle 1 to travel in the lateral direction. In this case, the wheel assembly 5 does not rotate about the axial center C2 thereof.

[0069] Further, if the rotating members 27R and 27L are rotatively driven in the same direction or opposite directions at velocities that are different from each other (velocities including directions), then the wheel assembly 5 will rotate about the axial center C2 and also rotate about the center C1 of the cross-section thereof.

[0070] At this time, motions combining the aforesaid rotational motions (combined motions) cause the wheel assembly 5 to travel in directions inclined relative to the fore-and-aft direction and the lateral direction, thus causing the whole vehicle 1 to travel in the same direction as that of the wheel assembly 5. The traveling direction of the wheel assembly 5 in this case will change, depending upon the difference between the rotational velocities, including the rotational directions, of the rotating members 27R and 27L (the rotational velocity vectors, the polarities of which are defined according to rotational directions).

[0071] The traveling motions of the wheel assembly 5 effected as described above. Therefore, by controlling the rotational velocities (including the rotational directions) of the electric motors 31R and 31L, and consequently by controlling the rotational velocities of the rotating members 27R and 27L, it becomes possible to control the moving velocity and the traveling direction of the vehicle 1.

[0072] Incidentally, the seat (the boarding portion) 3 and the base body 9 are tiltable about the lateral axial center C2, the axial center C2 of the wheel assembly 5 being the supporting point, and also tiltable together with the wheel assembly 5 about the longitudinal axis, the ground contact surface (the lower end surface) of the wheel assembly 5 being the supporting point.

[0073] (Construction of the Control Device of the Vehicle)

[0074] The construction for controlling the operation of the vehicle 1 according to the present embodiment will now be described. In the following description, assuming an XYZ coordinate system, in which the longitudinal horizontal axis is indicated by an X-axis, the lateral horizontal axis is indicated by a Y-axis, and the vertical direction is indicated by a Z-axis, as illustrated in FIG. 1 and FIG. 2, the fore-and-aft direction and the lateral direction may be referred to as the X-axis direction and the Y-axis direction, respectively.

[0075] For example, the fore-and-aft direction is defined as a “first direction”, and the lateral direction is defined as a “second direction”. The azimuth of the front direction or the right direction of the vehicle 1 in the world coordinate system is defined as an “orientation” of the vehicle 1.

[0076] First, the control of the operation of the vehicle 1 will be outlined. According to the present embodiment, basically, if the occupant seated on the seat 3 tilts his/her upper body (more specifically, if the upper body is tilted such that the position of the overall center-of-gravity point combining the occupant and the vehicle 1 (the position projected onto a horizontal plane) is moved), then the base body 9 is tilted together with the seat 3 toward the side to which the upper body has been tilted. At this time, the traveling motion of the wheel assembly 5 is controlled such that the vehicle 1 travels toward the side to which the base body 9 has tilted. For example, if the occupant tilts his/her upper body forward, causing the base body 9 to tilt forward together with the seat 3, then the traveling motion of the wheel assembly 5 is controlled to cause the vehicle 1 to travel forward.

[0077] In other words, according to the present embodiment, the operation in which the occupant moves his/her upper body, causing the seat 3 and the base body 9 to tilt provides one basic steering operation for the vehicle 1 (a motion request of the vehicle 1), and the traveling motion of the wheel assembly 5 is controlled through the actuator 7 according to the steering operation.

[0078] Here, in the vehicle 1 according to the present embodiment, the ground contact surface of the wheel assembly 5 as the ground contact surface of the whole vehicle 1 will be a single local region which is smaller than a region resulting from projecting all the vehicle 1 and the occupant thereon onto a floor surface, and a floor reaction force will act only on the single local region. For this reason, in order to prevent the base body 9 from falling due to tilting, the wheel assembly 5 must be moved such that the overall center-of-gravity point of the occupant and the vehicle 1 is positioned substantially right above the ground contact surface of the wheel assembly 5.

[0079] Therefore, according to the present embodiment, the posture of the base body 9 in a state wherein the overall center-of-gravity point of the occupant and the vehicle 1 is positioned substantially right above the central point of the wheel assembly 5 (the central point on the axial center C2) (more precisely, in a state wherein the center-of-gravity point is positioned substantially right above the ground contact surface of the wheel assembly 5) is defined as a desired posture, and basically, the traveling motion of the wheel assembly 5 is controlled such that the actual posture of the base body 9 is converged to the desired posture.

[0080] Further, in a state wherein no occupant is aboard the vehicle 1, the posture of the base body 9 in a state which the center-of-gravity point of the vehicle 1 alone is positioned substantially right above the central point of the wheel assembly 5 (the central point on the axial center C2) (more precisely, a state wherein the center-of-gravity point is positioned substantially right above the ground contact surface of the wheel assembly 5) is defined as a desired posture, and the actual posture of the base body 9 is converged to the desired posture. Thus, the traveling motion of the wheel assembly 5 is controlled such that the vehicle 1 supports itself without causing the base body 9 to fall from tilting.

[0081] Further, the traveling motion of the wheel assembly 5 is controlled so that, in either state of the state in which the occupant is aboard the vehicle 1 and in which the occupant is not aboard the vehicle 1, the moving velocity of the vehicle 1 increases as the deviation of the actual posture of the base body 9 from the desired posture increases, and the movement of the vehicle 1 stops in the case where the actual posture of the base body 9 coincides with the desired posture.

[0082] Supplementally, “the posture” means a spatial orientation. In the present embodiment, when the base body 9 tilts together with the seat 3, the postures of the base body 9 and the seat 3 change. Further, in the present embodiment, the base body 9 and the seat 3 integrally tilt, so that converging the posture of the base body 9 to the desired posture is equivalent to converging the posture of the seat 3 to a desired posture associated with the seat 3 (the posture of the seat 3 in a state wherein the posture of the base body 9 coincides with a desired posture of the base body 9).

[0083] According to the present embodiment, in order to control the operation of the vehicle 1 as described above, a control unit 50 constituted of an electronic circuit unit which mainly includes a microcomputer and a drive circuit unit for the electric motors 31R and 31L, a tilt sensor 52 for measuring a tilt angle θb relative to the vertical direction (the gravitational direction) of a predetermined portion of the base body 9 and a changing velocity thereof (=dθb/dt), a load sensor 54 for detecting whether or not an occupant is aboard the vehicle 1, and rotary encoders 56R and 56L serving as angle sensors for detecting the rotational angles and the rotational angular velocities of the output shafts of the electric motors 31R and 31L, respectively, are mounted at appropriate places of the vehicle 1, as illustrated in FIG. 1 and FIG. 2.

[0084] In this case, the control unit 50 and the tilt sensor 52 are installed to the support frame 13 by, for example, being accommodated in the support frame 13 of the base body 9. Further, the load sensor 54 is incorporated in the seat 3. Further, the rotary encoders 56R and 56L are provided integrally with the electric motors 31R and 31L. The rotary encoders 56R and 56L may alternatively be attached to the rotating members 27R and 27L, respectively.

[0085] More specifically, the aforesaid tilt sensor 52 is constructed of an acceleration sensor and a rate sensor (angular velocity sensor), such as a gyro sensor, and outputs detection signals of these sensors to the control unit 50. Then, the control unit 50 carries out predetermined measurement arithmetic processing (this may be publicly known arithmetic processing) on the basis of the outputs of the acceleration sensor and the rate sensor of the tilt sensor 52 thereby to calculate the measured value of the tilt angle θb of the portion, to which the tilt sensor 52 is installed (the support frame 13 in the present embodiment), relative to the vertical direction and the measured value of the tilt angular velocity θbdot, which is a change rate (differential value) thereof.

[0086] In this case, to be more specific, the tilt angle θb to be measured (hereinafter referred to a base body tilt angle θb in some cases) is constituted of a component in the direction about the Y-axis (a pitch direction) θb_x and a component in the direction about the X-axis (a roll direction) θb_y. Similarly, the tilt angular velocity θbdot to be measured (hereinafter referred to a base body tilt angular velocity θbdot in some cases) is constituted of a component in the direction about the Y-axis (the pitch direction) θbdot_x (=dθb_x/dt) and a component in the direction about the X-axis (the roll direction) θbdot_y (=dθb_y/dt).

[0087] Supplementally, according to the present embodiment, the seat 3 tilts integrally with the support frame 13 of the base body 9, so that the base body tilt angle θb also has a meaning as the tilt angle of the payload supporting part 3.

[0088] In the description of the present embodiment, regarding variables, such as a motional state amount having components in directions of the X-axis and the Y-axis, such as the aforesaid base body tilt angle θb (or directions about axes), or variables, such as coefficients related to the motional state amount, the reference characters of the variables will be accompanied by a suffix “_x” or “_y” to distinguishably denote the components.

[0089] In this case, for the variables related to translational motions, such as a translational velocity, a component in the X-axis direction thereof will be accompanied by the suffix “_x” and a component in the Y-axis direction thereof will be accompanied by the suffix “_y.”

[0090] Meanwhile, regarding the variables related to rotational motions, such as angles, rotational velocities (angular velocities), and angular acceleration, for the purpose of convenience, a component in the direction about the Y-axis will be accompanied by the suffix “_x” and a component in the direction about the X-axis will be accompanied by the suffix “_y” in order to match the variables related to translational motions with suffixes.

[0091] Further, to denote a variable in the form of a pair of a component in the X-axis direction (or a component in the direction about the Y-axis) and a component in the Y-axis direction (or a component in the direction about the X-axis), the suffix “_xy” is added to the reference character of the variable. For example, to express the aforesaid base body tilt angle θb in the form of the pair of a component in the direction about the Y-axis θb_x and a component in the direction about the X-axis θb_y, the pair will be denoted by “the base body tilt angle θb_xy.”

[0092] The load sensor 54 is incorporated in the seat 3 so as to be subjected to a load from the weight of an occupant when the occupant sits on the seat 3, and outputs a detection signal based on the load to the control unit 50. Then, the control unit 50 determines whether or not the occupant is aboard the vehicle 1 on the basis of the measured value of the load indicated by the output of the load sensor 54.

[0093] In place of the load sensor 54, a switch type sensor which, for example, turns on when an occupant sits on the seat 3 may be used.

[0094] The rotary encoder 56R generates a pulse signal each time the output shaft of the electric motor 31R rotates for a predetermined angle, and outputs the pulse signal to the control unit 50. Then, based on the pulse signal, the control unit 50 measures the rotational angle of the output shaft of the electric motor 53R and further measures the temporal change rate (differential value) of the measured value of the rotational angle as the rotational angular velocity of the electric motor 53R. The same applies to the rotary encoder 56L for the electric motor 31L.

[0095] The control unit 50 carries out predetermined arithmetic processing by using the aforesaid measured values thereby to determine velocity commands, which are the desired values of the rotational angular velocities of the electric motors 31R and 31L, respectively, and carries out feedback control on the rotational angular velocity of each of the electric motors 31R and 31L according to the determined velocity commands.

[0096] Incidentally, the relationship between the rotational angular velocity of the output shaft of the electric motor 31R and the rotational angular velocity of the rotating member 27R will be a proportional relationship based on the speed reduction ratio of a fixed value between the output shaft and the rotating member 27R. Hence, for the sake of convenience, in the description of the present embodiment, the rotational angular velocity of the electric motor 31R will mean the rotational angular velocity of the rotating member 27R. Similarly, the rotational angular velocity of the electric motor 31L will mean the rotational angular velocity of the rotating member 27L.

[0097] (Outline of the Control Method of the Vehicle)

[0098] The following will describe in more detail the control processing carried out by the control unit 50.

[0099] The control unit 50 executes the processing (main routine processing) illustrated by the flowchart of FIG. 7 at a predetermined control processing cycle.

[0100] First, in STEP1, the control unit 50 acquires an output of a tilt sensor 52.

[0101] Subsequently, the control unit 50 proceeds to STEP2 to calculate a measured value θb_xy_s of a base body tilt angle θb and a measured value θbdot_xy_s of a base body tilt angular velocity θbdot on the basis of the acquired output of the tilt sensor 52.

[0102] In the following description, to denote the observed value (the measured value or an estimated value) of an actual value of a variable (a state amount), such as the aforesaid measured value θb_xy_s of the base body tilt angle θb, by a reference character, the reference character of the variable will have a suffix “_s”.

[0103] Next, after acquiring an output of a load sensor 54 in STEP3, the control unit 50 carries out the determination processing in STEP4. In the determination processing, the control unit 50 determines whether or not the vehicle 1 has an occupant aboard (whether or not an occupant is sitting on the seat 3) by determining whether or not the load measured value indicated by the acquired output of the load sensor 54 is larger than a predetermined value which has been set beforehand.

[0104] Then, if the determination result in STEP4 is affirmative, then the control unit 50 carries out the processing for setting a desired value θb_xy_obj of the base body tilt angle θb and the processing for setting the values of constant parameters (e.g., the basic values of various gains) for controlling the operation of the vehicle 1 in STEP5 and STEP6, respectively.

[0105] In STEP5, the control unit 50 sets a predetermined desired value for a boarding mode as the desired value θb_xy_obj of the base body tilt angle θb.

[0106] Here, the term “boarding mode” means the operation mode of the vehicle 1 in the case where the occupant is aboard the vehicle 1. The desired value θb_xy_obj for the boarding mode is preset such that desired value θb_xy_obj coincides or substantially coincides with the measured value θb_xy_s of the base body tilt angle θb measured on the basis of an output of the tilt sensor 52 in a posture of the base body 9 in which the overall center-of-gravity point of the vehicle 1 and the occupant seated on the seat 3 (hereinafter referred to as the vehicle-occupant overall center-of-gravity point) is positioned substantially right above a ground contact surface of the wheel assembly 5.

[0107] Further, in STEP6, the control unit 50 sets predetermined values for the boarding mode as the values of constant parameters for controlling the operation of the vehicle 1. The constant parameters include, for example, hx, hy, Ki_a_x, Ki_b_x, Ki_a_y, and Ki_b_y (i=1, 2, 3), which will be discussed later.

[0108] Meanwhile, if the determination result in STEP4 is negative, then the control unit 50 carries out the processing for setting a desired value θb_xy_obj of a base body tilt angle θb_xy and the processing for setting the values of constant parameters for controlling the operation of the vehicle 1 in STEP7 and STEP8, respectively.

[0109] In STEP7, the control unit 50 sets a predetermined desired value for an autonomous mode as the desired value θb_xy_obj of the base body tilt angle θb.

[0110] Here, the term “autonomous mode” means an operation mode of the vehicle 1 in the case where the occupant is not aboard the vehicle 1. The desired value θb_xy_obj for the autonomous mode is preset such that desired value θb_xy_obj coincides or substantially coincides with the measured value θb_xy_s of the base body tilt angle θb measured on the basis of an output of the tilt sensor 52 in a posture of the base body 9 in which the center-of-gravity point of the vehicle 1 alone (hereinafter referred to as the vehicle-alone center-of-gravity point) is positioned substantially right above the ground contact surface of the wheel assembly 5. The desired value θb_xy_obj for the autonomous mode is generally different from the desired value θb_xy_obj for the boarding mode.

[0111] Further, in STEP8, the control unit 50 sets predetermined values for the autonomous mode as the values of constant parameters for controlling the operation of the vehicle 1. The values of the constant parameters for the autonomous mode are different from the values of the constant parameters for the boarding mode.

[0112] The aforesaid values of the constant parameters are set to be different between the boarding mode and the autonomous mode, because the response characteristics of the operations of the vehicle 1 relative to control inputs are different from each other due to the differences in the height of the aforesaid center-of-gravity point, the overall mass, and the like between the respective modes.

[0113] By the processing in STEP4 to STEP8 described above, the desired value θb_xy_obj of the base body tilt angle θb_xy and the values of the constant parameters are set for each of the operational modes, namely, the boarding mode and the autonomous mode.

[0114] Incidentally, the processing in STEP5 and STEP6 or the processing in STEP7 and STEP8 is not essential to carry out for each control processing cycle. Alternatively, the processing may be carried out only when the determination result in STEP4 changes.

[0115] Supplementally, in both the boarding mode and the autonomous mode, the desired value of a component θbdot_x in the direction about a Y-axis of the base body tilt angular velocity θbdot and the desired value of a component θbdot_y in the direction about an X-axis thereof are both zero. For this reason, it is unnecessary to carry out the processing for setting a desired value of the base body tilt angular velocity θbdot_xy.

[0116] After carrying out the processing in STEP5 and STEP6 or the processing in STEP7 and STEP8 as described above, the control unit 50 carries out vehicle control arithmetic processing in STEP9 thereby to determine the speed commands for electric motors 31R and 31L, respectively. The vehicle control arithmetic processing will be discussed later in detail.

[0117] Subsequently, the control unit 50 proceeds to STEP10 to carry out the processing for controlling the operations of the electric motors 31R and 31L according to the speed commands determined in STEP9. In this operation control processing, based on the difference between the speed command for the electric motor 31R determined in STEP9 and the measured value of the rotational speed of the electric motor 31R measured on the basis of an output of a rotary encoder 56R, the control unit 50 determines a desired value (desired torque) of an output torque of the electric motor 31R such that the difference is converged to zero. Then, the control unit 50 controls the current supplied to the electric motor 31R such that the electric motor 31R outputs an output torque of the desired torque. The same applies to the operation control of the left electric motor 31L.

[0118] The above has described the general control processing carried out by the control unit 50.

[0119] (Details of the Control Method of the Vehicle)

[0120] The vehicle control arithmetic processing in STEP9 mentioned above will now be described in detail.

[0121] In the following description, the vehicle-occupant overall center-of-gravity point in the boarding mode and the vehicle-alone center-of-gravity point in the autonomous mode will be generically referred to as the vehicle system center-of-gravity point. The vehicle system center-of-gravity point will mean the vehicle-occupant overall center-of-gravity point when the operational mode of the vehicle 1 is the boarding mode and will mean the vehicle-alone center-of-gravity point when the operational mode of the vehicle 1 is the autonomous mode.

[0122] Further, in the following description, regarding the values (updated values) determined at each control processing cycle by the control unit 50, a value determined at a current (latest) control processing cycle may be referred to as a current value, and a value determined at an immediately preceding control processing cycle may be referred to as a previous value. Further, a value will mean a current value unless otherwise specified as a current value or a previous value.

[0123] Further, regarding the velocity and acceleration in the X-axis direction, a forward-facing direction will be defined as a positive direction, and regarding the velocity and acceleration in the Y-axis direction, a left-facing direction will be defined as the positive direction.

[0124] In the present embodiment, the vehicle control arithmetic processing in STEP9 is carried out, assuming that the dynamic behaviors of the vehicle system center-of-gravity point (more specifically, the behaviors observed by projecting the behaviors from the Y-axis direction onto a plane (XZ plane) which is orthogonal thereto, and the behaviors observed by projecting the behaviors from the X-axis direction onto a plane (YZ plane) which is orthogonal thereto) are approximately expressed by the behaviors of an inverted pendulum model (dynamic behaviors of the inverted pendulum), as shown in FIG. 8.

[0125] In FIG. 8, unparenthesized reference numerals denote the reference numerals associated with the inverted pendulum model observed from the Y-axis direction, while the parenthesized reference numerals denote the reference numerals associated with the inverted pendulum model observed from the X-axis direction.

[0126] In this case, the inverted pendulum model expressing a behavior observed from the Y-axis direction is provided with a mass point 60_—x positioned at the vehicle system center-of-gravity point and a imaginary wheel 62_—x, which has a rotational axis 62a_—x parallel to the Y-axis direction and which freely circumrotate on a floor surface (hereinafter referred to as the imaginary wheel 62_—x). Further, the mass point 60_—x is supported by a rotational shaft 62a_—x of the imaginary wheel 62_—x through the intermediary of a linear rod 64_—x such that the mass point 60_—x is swingable about the rotational shaft 62a_—x, using the rotational shaft 62a_—x as the supporting point.

[0127] In this inverted pendulum model, a motion of the mass point 60_—x corresponds to a motion of the vehicle system center-of-gravity point observed from the Y-axis direction. Further, it is assumed that the tilt angle θbe_x of the rod 64_—x relative to a vertical direction coincides with a difference θbe_x_s between a base body tilt angle measured value θb_x_s in the direction about the Y-axis and a base body tilt angle desired value θb_x_obj (=θb_x_s−θb_x_obj). It is also assumed that a changing velocity of the tilt angle θbe_x of the rod 64_—x (=θbe_x/dt) coincides with a base body tilt angular velocity measured value θbdot_x_s in the direction about the Y-axis. Further, it is assumed that a moving velocity Vw_x of the imaginary wheel 62_—x (the translational moving velocity in the X-axis direction) coincides with the moving velocity of the wheel assembly 5 of the vehicle 1 in the X-axis direction.

[0128] Similarly, the inverted pendulum model expressing a behavior observed from the X-axis direction (refer to the parenthesized reference numerals in FIG. 8) is provided with a mass point 60_—y positioned at the vehicle system center-of-gravity point and a imaginary wheel 62_—y, which has a rotational axis 62a_—y parallel to the X-axis direction and which freely circumrotate on a floor surface (hereinafter referred to as the imaginary wheel 62_—y). Further, the mass point 60_—y is supported by a rotational shaft 62a_—y of the imaginary wheel 62_—y through the intermediary of a linear rod 64_—y such that the mass point 60_—y is swingable about the rotational shaft 62a_—y, using the rotational shaft 62a_—y as the supporting point.

[0129] In this inverted pendulum model, a motion of the mass point 60_—y corresponds to a motion of the vehicle system center-of-gravity point observed from the X-axis direction. Further, it is assumed that the tilt angle θbe_y of the rod 64_—y relative to the vertical direction coincides with a difference θbe_y_s between a base body tilt angle measured value θb_y_s in the direction about the X-axis and a base body tilt angle desired value θb_y_obj (=θb_y_s−θb_y_obj). It is also assumed that a changing velocity of the tilt angle θbe_y of the rod 64_—y (=dθbe_y/dt) coincides with a base body tilt angular velocity measured value θbdot_y_s in the direction about the X-axis. Further, it is assumed that a moving velocity Vw_y of the imaginary wheel 62_—y (the translational moving velocity in the Y-axis direction) coincides with the moving velocity of the wheel assembly 5 of the vehicle 1 in the Y-axis direction.

[0130] It is assumed that the imaginary wheels 62_—x and 62_—y have radii Rw_x and Rw_y of predetermined values, respectively.

[0131] It is assumed that relationships represented by expressions 01a and 01b given below hold between rotational angular velocities ωw_x and ωw_y of the imaginary wheels 62_—x and 62_—y, respectively, and rotational angular velocities ω_R and ω_L of the electric motors 31R and 31L, respectively (more accurately, the rotational angular velocities ω_R and ω_L of rotational members 27R and 27L, respectively).
ωw_—x=(ω_—R+ω_—L)/2  Expression 01a
ωw_—y=C·(ω_—R−ω_—L)/2  Expression 01b

[0132] Where “C” in expression 01b denotes a coefficient of a predetermined value that depends on a mechanical relationship or slippage between free rollers 29R and 29L and the wheel assembly 5. The positive directions of ωw_x, ω_R and ω_L are the directions in which the imaginary wheel 62_—x rotates in the case where the imaginary wheel 62_—x circumrotates forward. The positive direction of ωw_y is the direction in which the imaginary wheel 62_—y rotates in the case where the imaginary wheel 62_—y circumrotates leftwards.

[0133] Here, the dynamics of the inverted pendulum model shown in FIG. 8 is represented by expressions 03x and 03y given below. Expression 03x is an expression that represents the dynamics of the inverted pendulum model observed from the Y-axis direction, while expression 03y is an expression that represents the dynamics of the inverted pendulum model observed from the X-axis direction.
d²θbe_—x/dt²=α_—x·θbe_—x+β_—x·ωwdot_—x  Expression 03x
d²θbe_—y/dt²=α_—y·θbe_—y+β_—y·ωwdot_—y  Expression 03y

[0134] Where ωwdot_x in expression 03x denotes the rotational angular acceleration (first-order differential value of the rotational angular velocity ωw_x) of the imaginary wheel 62_—x, α_x denotes a coefficient which depends on a mass or a height h_x of the mass point 60_—x, and β_x denotes a coefficient which depends on an inertia (inertial moment) or the radius Rw_x of the imaginary wheel 62_—x. The same applies to ωwdot_y, α_y, and β_y in expression 03y.

[0135] As may be understood from these expressions 03x and 03y, the motions of the mass points 60_—x and 60_—y of the inverted pendulum model (i.e., the motions of the vehicle system center-of-gravity point) are specified, depending on the rotational angular acceleration ωwdot_x of the imaginary wheel 62_—x and the rotational angular acceleration ωwdot_y of the imaginary wheel 62_—y, respectively.

[0136] In the present embodiment, therefore, the rotational angular acceleration ωwdot_x of the imaginary wheel 62_—x is used as the manipulated variable (control input) for controlling the motion of the vehicle system center-of-gravity point observed from the Y-axis direction, while the rotational angular acceleration ωwdot_y of the imaginary wheel 62_—y is used as the manipulated variable (control input) for controlling the motion of the vehicle system center-of-gravity point observed from the X-axis direction.

[0137] To briefly describe the vehicle control arithmetic processing in STEP9, the control unit 50 determines imaginary wheel rotational angular acceleration commands ωwdot_x_cmd and ωwdot_y_cmd, which are the command values (desired values) of the rotational angular accelerations ωwdot_x and ωwdot_y as the manipulated variables such that the motion of the mass point 60_—x observed in the X-axis direction and the motion of the mass point 60_—y observed in the Y-axis direction become the motions corresponding to desired motions of the vehicle system center-of-gravity point. Further, the control unit 50 determines the values obtained by integrating the imaginary wheel rotational angular acceleration commands ωwdot_x_cmd and ωwdot_y_cmd, respectively, as the imaginary wheel rotational angular velocity commands ωw_x_cmd and ωw_y_cmd, which are the command values (desired values) of the rotational angular velocities ωw_x and ωw_y of the imaginary wheels 62_—x and 62_—y, respectively.

[0138] Further, the control unit 50 defines the moving velocity of the imaginary wheel 62_—x corresponding to the imaginary wheel rotational angular velocity command ωw_x_cmd (=Rw_x·ωw_x_cmd) and the moving velocity of the imaginary wheel 62_—y corresponding to the imaginary wheel rotational angular velocity command ωw_y_cmd (=Rw_y·ωw_y_cmd) as the desired moving velocity of the wheel assembly 5 of the vehicle 1 in the X-axis direction and the desired moving velocity thereof in the Y-axis direction, respectively, and the control unit 50 determines velocity commands ω_R_cmd and ω_L_cmd of the electric motors 31R and 31L, respectively, so as to achieve the desired moving velocities.

[0139] In the present embodiment, the imaginary wheel rotational angular acceleration commands ωwdot_x_cmd and ωwdot_y_cmd as the manipulated variables (control inputs) are determined by adding up three manipulated variable components, as indicated by expressions 07x and 07y, which will be discussed later.

[0140] Supplementally, of the imaginary wheel rotational angular acceleration commands ωwdot_x_cmd and ωwdot_y_cmd as the manipulated variable (control input) in the present embodiment, ωwdot_x_cmd is a rotational angular acceleration of the imaginary wheel 62_—x moving in the X-axis direction, so that it functions as the manipulated variable which stipulates the driving force to be imparted to the wheel assembly 5 so as to make the wheel assembly 5 move in the X-axis direction. Further, because ωwdot_y_cmd is a rotational angular acceleration of the imaginary wheel 62_—y moving in the Y-axis direction, so that it functions as the manipulated variable which stipulates the driving force to be imparted to the wheel assembly 5 so as to make the wheel assembly 5 move in the Y-axis direction.

[0141] The control unit 50 is provided with the functions illustrated in the block diagram of FIG. 17 as the functions for carrying out the vehicle control arithmetic processing in STEP9 as described above.

[0142] More specifically, the control unit 50 is provided with an error calculator 70 which calculates the base body tilt angle error measured value θbe_xy_s, which is the difference between the base body tilt angle measured value θb_xy_s and the base body tilt angle desired value θb_xy_obj, a center-of-gravity velocity calculator 72 which calculates an estimated center-of-gravity velocity value Vb_xy_s as an observed value of a center-of-gravity velocity Vb_xy, which is the moving velocity of the vehicle system center-of-gravity point, a center-of-gravity velocity restrictor 76 which determines a desired center-of-gravity velocity for control Vb_xy_mdfd as the desired value of the center-of-gravity velocity Vb_xy by taking into account a restriction based on a permissible range of the rotational angular velocities of the electric motors 31R and 31L, and a gain adjustor 78 which determines a gain adjustment parameter Kr_xy for adjusting the values of the gain coefficients of expressions 07x and 07y, which will be discussed later.

[0143] The control unit 50 is further provided with a posture control calculator 80 which calculates the imaginary wheel rotational angular velocity command ωw_xy_cmd, and a motor command calculator 82 which converts the imaginary wheel rotational angular velocity command ωw_xy_cmd into a pair of a velocity command ω_R_cmd (a command value of a rotational angular velocity) for the right electric motor 31R and a velocity command ω_L_cmd (a command value of a rotational angular velocity) for the left electric motor 31L.

[0144] Reference numeral 84 in FIG. 17 denotes a delay element which receives the imaginary wheel rotational angular velocity command ωw_xy_cmd calculated at each control processing cycle by the posture control calculator 80. The delay element 84 outputs a previous value ωw_xy_cmd_p of the imaginary wheel rotational angular velocity command ωw_xy_cmd at each control processing cycle.

[0145] In the vehicle control arithmetic processing in STEP9 described above, the processing by the aforesaid processing sections is carried out as described below.

[0146] The control unit 50 first carries out the processing by the error calculator 70 and the processing by the center-of-gravity velocity calculator 72.

[0147] The error calculator 70 receives the base body tilt angle measured values θb_xy_s (θb_x_s and θb_y_s) calculated in the aforesaid STEP2 and the desired values θb_xy_obj (θb_x_obj and θb_y_obj) set in the aforesaid STEP5 or STEP7. Then, the error calculator 70 subtracts θb_x_obj from θb_x_s to calculate the base body tilt angle error measured value θbe_x_s (=θb_x_s−θb_x_obj) in the direction about the Y-axis, and also subtracts θb_y_obj from θb_y_s to calculate the base body tilt angle error measured value θbe_y_s (=θb_y_s−θb_y_obj) in the direction about the X-axis.

[0148] The processing by the error calculator 70 may be carried out before the vehicle control arithmetic processing in STEP9. For example, the processing by the error calculator 70 may be carried out during the processing in the aforesaid STEP5 or STEP7.

[0149] Explanation will be given on the setting method of setting the desired value θb_xy_obj of the base body tilt angle θb in each of STEP 5 and STEP 7. In setting the desired value θb_xy_obj of the base body tilt angle θb, (1) an estimated value θb_xy_s_world of the base body tilt angle θb in a world coordinate system, and (2) an estimated value θslope_xy_s of a floor surface inclination angle θslope with respect to the horizontal surface in a vehicle coordinate system, are calculated.

[0150] The suffix “_world” means that the value belongs to the world coordinate system, and on the other hand, the absence of this suffix means that the value belongs to the vehicle coordinate system. Further, the suffix “_xyz” collectively represents the component about the Y-axis, the component about the X-axis, and the component about the Z-axis, or the component in the X-axis direction, the component in the Y-axis direction, and the component in the Z-axis direction.

[0151] (Method of Estimating the Base Body Tilt Angle)

[0152] First, the method of calculating the base body tilt angle (estimated value) θb_xyz_s_world will be explained with reference to FIG. 9. The construction shown in FIG. 9 (especially the construction for calculating the Z component θb_z_s_world) corresponds to a “vehicle orientation estimating component”. In calculating the base body tilt angle θb_xyz_s_world, a base body tilt angular velocity (measured value) ωb_xyz_s about each of the Y-axis, X-axis, and Z-axis, respectively, and a base body acceleration (measured value) αb_xyz_s in each of the X-axis direction, Y-axis direction, and Z-axis direction, respectively, are used.

[0153] Each of the base body tilt angular velocity ωb_x_s about the X-axis, the base body tilt angular velocity ωb_y_s about the Y-axis, and a base body rotational angular velocity about the Z-axis (hereinafter referred to as “the base body tilt angular velocity”) ωb_z_s, respectively, are calculated on the basis of the output signals from the tilt sensor 52 (to be more precise, a three-axis velocity sensor configuring the tilt sensor 52). The base body acceleration αb_xyz_s in each of the X-axis direction, Y-axis direction, and Z-axis direction (perpendicular direction) are calculated on the basis of the output signals from the tilt sensor 52 (to be more precise, the three-axis velocity sensor configuring the tilt sensor 52).

[0154] It may be set so that translational vectors representing the position in the vehicle coordinate system of the tilt sensor 52, or the acceleration sensor and the angular velocity sensor configuring the tilt sensor 52 are stored in a memory, and the translational vectors may be used in estimating the base body tilt angular velocity ωb_xyz_s about each of the Y-axis, X-axis, and the Z-axis in the vehicle coordinate system, and the base body acceleration αb_xyz_s in each of the X-axis direction, Y-axis direction, and Z-axis direction in the vehicle coordinate system.

[0155] Similarly, it may be set so that rotation matrixes representing the posture in the vehicle coordinate system of the tilt sensor 52, or the acceleration sensor and the angular velocity sensor configuring the tilt sensor 52 are stored in the memory, and the rotation matrixes may be used in estimating the base body tilt angular velocity ωb_xyz_s about each of the Y-axis, X-axis, and the Z-axis in the vehicle coordinate system, and the base body acceleration αb_xyz_s in each of the X-axis direction, Y-axis direction, and Z-axis direction in the vehicle coordinate system.

[0156] A vehicle-world coordinate converter 221 calculates the base body acceleration αb_xyz_s_world in the world coordinate system, on the basis of the base body acceleration αb_xyz_s in the vehicle coordinate system. A coordinate transformation matrix from the vehicle coordinate system to the world coordinate system (or a quaternion equivalent thereto) used in the vehicle-world coordinate converter 221 is calculated as an inverse matrix of the rotation matrix taking the result obtained by passing the base body tilt angle θb_xyz_s_world in the world coordinate system through a delay element 220 (a past value or the previous value θb_xyz_s_p_world of the base body tilt angle) as an Euler angle.

[0157] For example, the result of the conversion of a vector ^t(αb_x_s, αb_y_s, αb_z_s) representing the base body acceleration αb_xyz_s in the vehicle coordinate system by the vehicle-world coordinate transformation matrix is calculated as a vector ^t(αb_x_s_world, αb_y_s_world, αb_z_s_world) taking each components of the base body acceleration αb_xyz_s_world in the world coordinate system as the elements.

[0158] Further, in a calculator 222, the base body acceleration αb_xyz_s_world in the world coordinate system is multiplied by an inverse of an acceleration of gravity (1/g). Further, in a calculator 223, a result of inputting the calculation result of the calculator 222 to an arctan function arctan(αb_xy_s_world/g) is calculated as an acceleration tangent angle θacc_xyz. However, zero is output by the calculator 223 as the Z component θacc_zof the acceleration tangent angle. As the base body acceleration αb_xy_s_world becomes larger, the acceleration tangent angle θacc_xy becomes larger (refer to FIG. 16(a)). A calculator 224 calculates the difference between the base body tilt angle θb_xyz_s_world in the world coordinate system and the acceleration tangent angle θacc_xyz.

[0159] A vehicle-world coordinate converter 211 calculates the base body tilt angular velocity ωb_xyz_s_world in the world coordinate system, on the basis of the base body tilt angular velocity ωb_xyz_s in the vehicle coordinate system. The coordinate transformation matrix from the vehicle coordinate system to the world coordinate system used in the vehicle-world coordinate converter 211 is the same as the coordinate transformation matrix used in the vehicle-world coordinate converter 221.

[0160] For example, the result of the conversion of a vector ^t(ωb_x_s, ωb_y_s, ωb_z_s) representing the base body tilt angular velocity ωb_xyz_s in the vehicle coordinate system by the vehicle-world coordinate transformation matrix is calculated as a vector ^t(ωb_x_s_world, ωb_y_s_world, ωb_z_s_world) taking each component of the base body tilt angular velocity ωb_xyz_s_world in the world coordinate system as the elements.

[0161] In a calculator 215, the result of calculation by the calculator 224 (θb_xyz_s_world−θacc_xyz) is multiplied by a coefficient Ksd2. In a calculator 212, a result of passing the result of calculation by the calculator 215 Ksd2(θb_xyz_s_world−θacc_xyz) through an integrator component 216 (Ksd2/s) (θb_xyz_s_world−θacc_xyz) is subtracted from the base body tilt angular velocity ωb_xyz_s_world.

[0162] In a calculator 217, the result of calculation by the calculator 224 (θb_xyz_s_world−θacc_xyz) is multiplied by a coefficient Ksd1. In a calculator 213, the result of calculation by the calculator 212 ωb_xyz_s_world−(Ksd2/s) (θb_xyz_s_world−θacc_xyz) is subtracted from the result of calculation by the calculator 217 Ksd1(θb_xyz_s_world−θacc_xyz). Thereafter, by passing through the result of calculation by the calculator 213 through an integrator component 214, an estimated body tilt angle value θb_xyz_s_world in the world coordinate system is calculated.

[0163] That is, the estimated body tilt angle value θb_xyz_s_world is calculated according to each of relational expressions 11a through 11c.
θb_—x_—s_world=(1/s){ωb_—x_—s_world−(Ksd2/s)(θb_—x_—s_world−θacc_—x)−Ksd1(θb_—x_—s_world−θacc_—x)}  Expression 11a
θb_—y_—s_world=(1/s){(ωb_—y_—s_world−(Ksd2/s)(θb_—y_—s_world−θacc_—y)−Ksd1(θby_—s_world−θacc_—y)}  Expression 11b
θb_—z_—s_world=(1/s){(θb_—z_—s_world−(Ksd2/s)θb_—z_—s_world−Ksd1(θb_—z_—s_world)}  Expression 11c

[0164] As is apparent from the relational expression 11a and the relational expression 11b, the estimated body tilt angle value θb_xy_s_world is calculated by integrating the result of correcting the base body tilt angular velocity ωb_xy_s_world according to the difference between the base body tilt angle θb_xy_s_world and the acceleration tangent angle θacc_xy. This is because the changing mode of the base body tilt angle θb_xy_world varies according to the base body acceleration αb_xy_world (refer to FIG. 16(a)).

[0165] (Method of Estimating the Two-Dimensional Floor Inclination)

[0166] Next, the method of calculating an estimated value θslope_xy_s of a floor surface inclination angle θslope_xy (the two-dimensional floor inclination) will be explained with reference to FIG. 10 through FIG. 15. The configuration shown in FIG. 10 and either one of FIG. 11 through FIG. 15 corresponds to “the two-dimensional floor inclination estimation component”. The configuration shown in FIG. 10 corresponds to “the instantaneous provisional two-dimensional floor inclination calculation component”. The configuration shown in each of FIG. 11 through FIG. 15 corresponds to “the filter component”. After calculating an instantaneous provisional value θslope_xy_c of the inclination angle θslope_xy of the floor surface, an estimated value θslope_xy_s of the inclination angle of the floor surface is calculated. The affix “_c” represents the instantaneous provisional value (or a provisional value).

[0167] (Method of Calculating the Instantaneous Provisional Value of the Two-Dimensional Floor Inclination)

[0168] Next, the method of calculating the instantaneous provisional value θslope_xy_c of the floor surface inclination angle θslope_xy will be explained with reference to FIG. 10.

[0169] In a calculator 232, a coefficient Kt is multiplied to a measured value Imot_RL_s of the drive current of the motors 31R, 31L. The affix “RL” collectively represents the left and right components. Calculation results Kt·Imot_R_s and Kt·Imot_L_s by the calculator 232 correspond to estimated torque values Trq_R_s and Trq_L_s about the X-axis and about the Y-axis of the wheel assembly 5 in the vehicle coordinate system, respectively.

[0170] It may be set so that a correction factor according to sliding generated at the traveling motion unit 5 is calculated on the basis of either one or both of a current measured value Imot_RL_s or the rotational speed of the motors 31R, 31L, and according to a table or a conversion equation stored in the memory, and the estimated torque value Trq_RL_s may be corrected by subtracting the correction factor from the estimated torque value Trq_RL_s. Further, the estimated torque value Trq_RL_s may be corrected by multiplying to the estimated torque value Trq_RL_s a predetermined correction factor according to a viscosity resistance of the motors 31R and 31L.

[0171] Thereafter, the calculation results Trq_R_s and Trq_L_s by the calculator 232 is converted by an RL-XY converter 234 to the torque Trqw_x_s of the imaginary wheel 62_—x and the torque Trqw_y_s of the imaginary wheel 62_—y, according to relational expressions 12a and 12b. In the present embodiment, each of the torque Trqw_x_s of the imaginary wheel 62_—x and the torque Trqw_y_s of the imaginary wheel 62_—y correspond to the “motional state amount” of the vehicle 1.
Trqw_—x=(Trq_—R+Trq_—L)/2  Expression 12a
Trqw_—y=C·(Trq_—R−Trq_—L)/2  Expression 12b

[0172] In the case where the calculation result of the calculator 232 corresponds to the rotational angular velocities ω_R_s and ω_L_s of the motors 31R, 31L, respectively, the calculation result ω_RL_s by the calculator 232 may be converted to the torque Trqw_x_s of the imaginary wheel 62_—x and the torque Trqw_y_s of the imaginary wheel 62_—y by the RL-XY converter 234 according to the procedure explained hereinbelow.

[0173] First, a simultaneous equation obtained by replacing each of ωw_x, ωw_y, ω_R and ω_L in relational expressions 01a, 01b with each of ωw_x_s, ωw_y_s, ω_R_s and ω_L_s, respectively, is solved taking the ωw_x_s and ωw_y_s as unknowns. Further, by passing the solutions ωw_x_s and ωw_y_s of the simultaneous expressions, respectively, through a differentiating component, the rotational angular accelerations ωwdot_x and ωwdot_y of the imaginary wheels 62_—x and 62_—y are calculated, respectively. Thereafter, by multiplying each of the rotational angular accelerations ωwdot_x and ωwdot_y of the imaginary wheels 62_—x and 62_—y, respectively, with moments of inertia Iwx and Iwy (stored in the memory) of the imaginary wheels 62_—x, 62_—y, respectively, the torque Trqw_x_s of the imaginary wheel 62_—x and the torque Trqw_y_s of the imaginary wheel 62_—y are calculated.

[0174] Thereafter, by dividing the torque Trq_xy of the imaginary wheel 62_—xy by mg (m is a mass of the vehicle 1, and g is a gravitational acceleration) at a calculator 236, the instantaneous provisional value θslope_xy_c of the floor surface inclination angle in the vehicle coordinate system is calculated. The calculation by the calculator 236 is expressed by equations of motion 13a, 13b which include a translational force the vehicle 1 receives from the floor surface by the imaginary wheel 62_—xy (=(Trq_x/Rw_xy)/cos θslope_xy), and a gravitational component acting on the vehicle 1 in a direction parallel to the floor surface (≈mgθslope_xy/cos θslope_xy), in the inverted pendulum model shown in FIG. 8.
0=Trqw_—x/Rw_—x−mg·sin θslope_—x≈Trqw_—x/Rw_—x−mgθslope_—x, where αb_—x=0  Equation 13a
0=Trqw_—y/Rw_—y−mg·sin θslope_—y≈Trqw_—y/Rw_—y−mgθslope_—y, where αb_—y=0  Equation 13b

[0175] The instantaneous provisional value θslope_xy_c of the floor surface inclination angle may be calculated by the calculator 236 according to equations of motion 14a, 14b which includes a static friction (=μ_xymg·cos θslope_xy/cos θslope_xy, μ_xy is a coefficient of stastic friction between the floor surface and the imaginary wheel 62_—xy) in the force received by the vehicle 1 from the floor surface via the imaginary wheel 62_—xy. Of the reference “±” in the equations of motion 14a, 14b, “+” represents a force acting in a direction climbing an inclined floor surface, whereas “−” represents a force acting in a direction descending the inclined floor surface (the same applies hereinunder). In the present embodiment, “the force component acting on the vehicle 1 in a direction parallel to the floor surface” includes a gravity component −mg·sin θslope_xy, and a static friction μ′_ymg·cos θslope_xy.
0=±(Trq_—x/Rw_—x−p_—xmg·cos θslope_—x)−mg·sin θslope_—x  Equation 14a
0=±(Trq_—y/Rw_—y−p_—ymg·cos θslope_—y)−mg·sin θslope_—y  Equation 14b

[0176] The instantaneous provisional value θslope_xy_c of the floor surface inclination angle may be calculated by the calculator 236 according to equations of motion 15a, 15b, which includes an inertial force mαb_xy acting on the vehicle 1, in addition to the force the vehicle 1 receives from the floor surface via the imaginary wheel 62_—xy (=(Trq_x/Rw_xy)/cos θslope_xy) and the gravity component acting on the vehicle 1 in a direction parallel to the floor surface (≈mgθslope_xy/cos θslope_xy (an approximate equation based on a presumption that θslope_xy is small). In the present embodiment, “the motional state amount” includes accelerations αb_x and αb_y of the vehicle 1, in addition to the torque Trqw_x_s of the imaginary wheel 62_—x and the torque Trqw_y_s of the imaginary wheel 62_—y. As is explained before, the accelerations αb_x and αb_y of the vehicle 1 is calculated on the basis of the output signal from the acceleration sensor constituting the tilt sensor 52.
mαb_—x=±Trq_—x/Rw_—x−mgθslope_—x  Equation 15a
mαb_—y=±Trq_—y/Rw_—y−mgθslope_—y  Equation 15b

[0177] Further, the instantaneous provisional value θslope_xy_c of the floor surface inclination angle may be calculated by the calculator 236 according to equations of motion 16a and 16b, which includes a kinetic friction force (=μ′_xymg·cos θslope_xy/cos θslope_xy, μ′_xy is a kinetic friction coefficient between the floor surface and the imaginary wheel 62_—xy) in the force the vehicle 1 receives from the floor surface via the imaginary wheel 62_—xy. In the present embodiment, “the force component acting on the vehicle 1 in a direction parallel to the floor surface” includes the gravitational component −mg·sin θslope_xy and the kinetic friction force μ′_xymg·cos θslope_xy.
mαb_—x=±(Trq_—x/Rw_—x−μ′_—xmg·cos θslope_—x)−mg·sin θslope_—x  Equation 16a
mαb_—y=±(Trq_—y/Rw_—y−μ′_—ymg·cos θslope_—y)−mg·sin θslope_—y  Equation 16b

[0178] Next, the method of calculating the estimated floor surface inclination angle value θslope_xy_s will be explained with reference to FIG. 11 through FIG. 15.

[0179] (Method of Calculating the Estimated Value of the Two-Dimensional Floor Inclination (No. 1))

[0180] Explanation will be given on one embodiment of the method of calculating the estimated floor surface inclination angle value θslope_xy_s represented by the block diagram in FIG. 11.

[0181] First, the instantaneous provisional value θslope_xy_c of the floor surface inclination angle in the vehicle coordinate system is converted to a value in the world coordinate system in a vehicle-world coordinate converter 242. In the coordinate conversion in the vehicle-world coordinate converter 242, an inverse matrix of the rotation matrix taking the previous estimated value of the base body tilt angle (Z component) θb_z_s_p_world as an Euler angle, or a quaternion equivalent thereto, is used.

[0182] For example, after a vector ^t(cos θslope_x_c, 0, sin θslope_x_c) representing the instantaneous provisional value θslope_x_c of the floor surface inclination angle in the X-axis direction in the vehicle coordinate system is converted by the vehicle-world coordinate transformation matrix, the Z component of the conversion result is input to an arcsin function, and the result thereof is calculated as the estimated value θslope_x_s_world of the floor surface inclination angle in the X-axis direction in the world coordinate system. Similarly, after a vector ^t(0, cos θslope_y_c, sin θslope_y_c) representing the instantaneous provisional value θslope_y_c of the floor surface inclination angle in the Y-axis direction in the vehicle coordinate system is converted by the vehicle-world coordinate transformation matrix, the Z component of the conversion result is input to an arcsin function, and the result thereof is calculated as the estimated value θslope_y_s_world of the floor surface inclination angle in the Y-axis direction in the world coordinate system.

[0183] The result of passing the calculation result of the vehicle-world coordinate converter 242 through a low-pass filter 244 is calculated as the estimated floor surface inclination angle value θslope_xy_s_world in the world coordinate system. A transfer function of the low-pass filter 244 is represented by 1/(1+τs).

[0184] Thereafter, the estimated floor surface inclination angle value θslope_xy_s_world in the world coordinate system is converted to a value in the vehicle coordinate system in a world-vehicle coordinate converter 246. By doing so, the estimated value θslope_xy_s of the floor surface inclination angle in the vehicle coordinate system is calculated. In the world-vehicle coordinate converter 246, the inverse matrix of the coordinate transformation matrix or the quaternion equivalent thereto, that are used in the vehicle-world coordinate converter 242, are used.

[0185] For example, after a vector ^t(cos θslope_x_s_world, 0, sin θslope_x_s_world) representing the estimated floor surface inclination angle value θslope_x_s_world in the X-axis direction in the world coordinate system is converted by the world-vehicle coordinate transformation matrix, the Z component of the conversion result is input to an arcsin function, and the result thereof is calculated as the estimated floor surface inclination angle value θslope_x_s in the X-axis direction in the vehicle coordinate system. Similarly, after a vector ^t(0, cos θslope_y_s_world, sin θslope_y_s_world) representing the estimated floor surface inclination angle value θslope_y_s_world in the Y-axis direction of the world coordinate system is converted by the world-vehicle coordinate transformation matrix, the Z component of the conversion result is input to an arcsin function, and the result thereof is calculated as the estimated floor surface inclination angle value θslope_y_s in the Y-axis direction in the vehicle coordinate system.

[0186] (Method of Calculating the Estimated Value of the Two-Dimensional Floor Inclination (No. 2))

[0187] Explanation will be given on an embodiment of the method of calculating the estimated floor surface inclination angle value θslope_xy_s represented by the block diagram in FIG. 12.

[0188] The result of passing the estimated value θslope_xy_s of the floor surface inclination angle through a delay element 254 (a previous estimated value θslope_xy_s_p of the floor surface inclination angle in a current vehicle coordinate system) is converted to a value in the current vehicle coordinate system, in a previous-current vehicle coordinate converter 255. The calculation result of the previous-current vehicle coordinate converter 255 corresponds to an estimated reference value θslope_xy_s0 of the floor surface inclination angle in the current vehicle coordinate system.

[0189] In the previous-current vehicle coordinate converter 255, the rotation matrix taking the difference between a current value θb_z_s_world of the base body tilt angle (Z component) calculated by a calculator 256, and a previous value θb_z_s_world_p of the base body tilt angle (Z component), or a quaternion equivalent thereto, is used.

[0190] For example, a vector ^t(cos θslope_x_s_p, 0, sin θslope_x_s_p) representing the previous estimated value θslope_x_s_p of the floor surface inclination angle in the X-axis direction in the previous vehicle coordinate system is converted by the previous-current vehicle coordinate transformation matrix, the Z component of the conversion result is input to an arcsin function, and the result thereof is calculated as the estimated reference value θslope_x_s0 of the floor surface inclination angle in the X-axis direction in the current vehicle coordinate system. Similarly, a vector ^t(0, cos θslope_y_s_p, sin θslope_x_s_p) representing a previous estimated value θslope_y_s_p of the floor surface inclination angle in the Y-axis direction of the previous vehicle coordinate system is converted by the previous-current vehicle coordinate transformation matrix, the Z component of the conversion result is input to an arcsin function, and the result thereof is calculated as the estimated reference value θslope_y_s0 of the floor surface inclination angle in the Y-axis direction in the current vehicle coordinate system.

[0191] Further, in a calculator 251, the estimated reference value θslope_xy_s0 of the floor surface inclination angle in the current vehicle coordinate system is subtracted from the instantaneous provisional value θslope_xy_c of the floor surface inclination angle. Further, in a calculator 252, the calculation result of the calculator 251 θslope_xy_c−θslope_xy_s0 is multiplied by a coefficient K.

[0192] Thereafter, in a calculator 253, the calculation result of the calculator 252 K(θslope_xy_c−θslope_xy_s0) is added with the estimated reference value θslope_xy_s0 of the floor surface inclination angle in the current vehicle coordinate system. The calculation result of the calculator 253 Kθslope_xy_c+(1−K)θslope_xy_s0 is calculated as an estimiated value θslope_xy_s of the floor surface inclination angle in the current vehicle coordinate system.

[0193] According to the method of calculation represented by respective block diagram in FIG. 11 and FIG. 12, the estimated floor inclination angle value θslope_xy_s is calculated using a first-order low-pass filter. This is equivalent to calculating a weighed mean value of the instantaneous provisional value θslope_xy_c of the floor surface inclination angle, and the previous estimated value θslope_xy_s_p of the floor surface inclination angle.

[0194] (Method of Calculating the Estimated Value of the Two-Dimensional Floor Inclination (No. 3))

[0195] Explanation will be given on an embodiment of the method of calculating the estimated floor surface inclination angle value θslope_xy_s represented by the block diagram of FIG. 13.

[0196] First, the instantaneous provisional value θslope_xy_c of the floor surface inclination angle θslope_xy in the vehicle coordinate system is converted to a value in the world coordinate system in a vehicle-world coordinate converter 262. In the coordinate conversion in the vehicle-world coordinate converter 262, an inverse matrix of the rotation matrix taking the previous estimated value of the base body tilt angle (Z component) θb_z_s_p_world as an Euler angle, or a quaternion equivalent thereto, is used.

[0197] The calculation result of the vehicle-world coordinate converter 262 is sequentially passed through a first low-pass filter 264 and a second low-pass filter 266, and the result thereof is calculated as the estimated floor surface inclination angle value θslope_xy_s_world in the world coordinate system. The transfer function of the first low-pass filter 264 is represented by 1/(1+τ₁s). The transfer function of the second low-pass filter 266 is represented by 1/(1+τ₂s).

[0198] Thereafter, the estimated floor surface inclination angle value θslope_xy_s_world in the world coordinate system is converted to a value in the vehicle coordinate system in the world-vehicle coordinate converter 268. By doing so, an estimated value θslope_xy_s of the floor surface inclination angle in the vehicle coordinate system is calculated. In the world-vehicle coordinate converter 246, the inverse matrix of the coordinate transformation matrix or the quaternion equivalent thereto, that are used in the vehicle-world coordinate converter 242, is used.

[0199] (Method of Calculating the Estimated Value of the Two-Dimensional Floor Inclination (No. 4))

[0200] Explanation will be given on one embodiment of the method of calculating the estimated floor surface inclination angle value θslope_xy_s represented by the block diagram in FIG. 14.

[0201] The result of passing a primary estimated value θslope_xy_s1 of the floor surface inclination angle (an output value of a calculator 313 as “a first phase output value”) through a delay element 314 (a previous primary estimated value θslope_xy_s1_p of the floor surface inclination angle in the current vehicle coordinate system) is converted to a value in the current vehicle coordinate system in a previous-current vehicle coordinate converter 315. The calculation result of the previous-current vehicle coordinate converter 315 corresponds to an first reference estimate value θslope_xy_s01 of the floor surface inclination angle in the current vehicle coordinate system.

[0202] In the previous-current vehicle coordinate converter 315, a coordinate conversion matrix taking a difference between a current value of the base body tilt angle (Z component) θb_z_s_world calculated at a calculator 316 and the previous value of the base body tilt angle (Z component) θb_z_s_world_p as an Euler angle, or a quaternion equivalent thereto, is used.

[0203] Further, in a calculator 311, the first reference estimate value θslope_xy_s01 of the floor surface inclination angle in the current vehicle coordinate system is subtracted from the instantaneous provisional value θslope_xy_c of the floor surface inclination angle. Further, in a calculator 312, the calculation result of the calculator 311 θslope_xy_c−θslope_xy_s01 is multiplied by a first coefficient W1.

[0204] Still further, in the calculator 313, the calculation result of the calculator 312 W1 (θslope_xy_c−θslope_xy_s01) is added with the first reference estimate value θslope_xy_s01 of the floor surface inclination angle in the current vehicle coordinate system. The calculation result of the calculator 313 W1θslope_xy_c+(1−W1)θslope_xy_s01 is calculated as the primary estimated value θslope_xy_s1 (the first phase output value) of the floor surface inclination angle.

[0205] Further, the result of passing a secondary estimated value θslope_xy_s2 of the floor surface inclination angle (an output value of a calculator 323 as “a second phase output value”) through a delay element 324 (previous secondary estimated value θslope_xy_s2_p of the floor surface inclination angle in the current vehicle coordinate system) is converted to a value in the current vehicle coordinate system in a previous-current vehicle coordinate converter 325. The calculation result of the previous-current vehicle coordinate converter 325 corresponds to an second reference estimate value θslope_xy_s02 of the floor surface inclination angle in the current vehicle coordinate system.

[0206] In the previous-current vehicle coordinate converter 325, a coordinate conversion matrix taking a difference between a current value of the base body tilt angle (Z component) θb_z_s_world calculated at a calculator 326 and the previous value of the base body tilt angle (Z component) θb_z_s_world_p as an Euler angle, or a quaternion equivalent thereto, is used.

[0207] Further, in a calculator 321, the second reference estimate value θslope_xy_s02 of the floor surface inclination angle in the current vehicle coordinate system is subtracted from the primary estimated value θslope_xy_s1. Further, in a calculator 322, the calculation result of the calculator 321 θslope_xy_s1−θslope_xy_s02 is multiplied by a second coefficient W2.

[0208] Thereafter, the calculation result of the calculator 322 W2(θslope_xy_s1−θslope_xy_s02) is added with the second reference estimate value θslope_xy_s02 of the floor surface inclination angle in the current vehicle coordinate system in the calculator 323. The calculation result of the calculator 323 W2θslope_xy_s1+(1−W2) θslope_xy_s02 is calculated as the secondary estimated value θslope_xy_s2 of the floor surface inclination angle in the current vehicle coordinate system (=an estimated value θslope_xy_s of the floor surface inclination angle).

[0209] (Method of Calculating the Estimated Value of the Two-Dimensional Floor Inclination (No. 5))

[0210] Explanation will be given on the method of calculating the estimated floor surface inclination angle value θslope_xy_s represented by the block diagram in FIG. 15.

[0211] A current instantaneous provisional value θslope_xy_c of the floor surface inclination angle is converted to the value of the previous vehicle coordinate system in a current-previous vehicle coordinate converter 335. The output value of the current-previous vehicle coordinate converter 335 corresponds to the first reference estimate value θslope_xy_s01 of the floor surface inclination angle in the previous vehicle coordinate system.

[0212] In the current-previous vehicle coordinate converter 335, a coordinate conversion matrix taking a difference between a current value of the base body tilt angle (Z component) θb_z_s_world calculated at a calculator 336 and the previous value of the base body tilt angle (Z component) θb_z_s_world_p (this is different in polarity from an output value of a calculator 346 to be explained later) as an Euler angle, or a quaternion equivalent thereto, is used.

[0213] Further, by passing the primary estimated value θslope_xy_s_p1 of the floor surface inclination angle in the previous vehicle coordinate system (an output value of a calculator 333 as “the first phase output value”) through a delay element 334, the previous primary estimated value θslope_xy_s1_p of the floor surface inclination angle in the previous vehicle coordinate system is calculated. In a calculator 331, the previous primary estimated value θslope_xy_s1_p of the floor surface inclination angle in the previous vehicle coordinate system is subtracted from the previous first reference estimate value θslope_xy_s01 of the floor surface inclination angle in the previous vehicle coordinate system. Further, in a calculator 332, the calculation result of the calculator 331θslope_xy_s01−θslope_xy_s1_p is multiplied by the first coefficient W1.

[0214] Further, the calculation result of the calculator 332 W1(θslope_xy_s01−θslope_xy_s1_p) is added with the previous primary estimated value θslope_xy_s1_p of the floor surface inclination angle in the previous vehicle coordinate system in a calculator 333. Further, the calculation result of the calculator 333 W1θslope_xy_s01+(1−W1)θslope_xy_s1_p is calculated as the current primary estimated value θslope_xy_s1 of the floor surface inclination angle in the previous vehicle coordinate system.

[0215] By passing the secondary estimated value θslope_xy_s2 of the floor surface inclination angle in the previous vehicle coordinate system through a delay element 344, a previous secondary estimated value θslope_xy_s2_p of the floor surface inclination angle in the previous vehicle coordinate system is calculated. The previous secondary estimated value θslope_xy_s2_p of the floor surface inclination angle in the previous vehicle coordinate system is subtracted from the current primary estimated value θslope_xy_s1 of the floor surface inclination angle in the previous vehicle coordinate system in a calculator 341. Further, the calculation result of the calculator 341 θslope_xy_s1−θslope_xy_s2_p is multiplied by the second coefficient W2 in a calculator 342.

[0216] Further, the calculation result of the calculator 342 W2(θslope_xy_s1−θslope_xy_s2_p) is added with the previous secondary estimated value θslope_xy_s2_p of the floor surface inclination angle in the previous vehicle coordinate system, in a calculator 343. The calculation result of the calculator 343 W2θslope_xy_s1+(1−W2)θslope_xy_s2_p is calculated as the secondary estimated value θslope_xy_s2 of the floor surface inclination angle in the previous vehicle coordinate system.

[0217] Lastly, the secondary estimated value θslope_xy_s2 of the floor surface inclination angle in the previous vehicle coordinate system is converted to a value in the current vehicle coordinate system in a previous-current vehicle coordinate converter 345. In the previous-current vehicle coordinate converter 345, the coordinate transformation matrix taking the difference between the current value of the base body inclination angle (Z component) θb_z_s_world calculated by the calculator 346 and the previous value of the base body inclination angle (Z component) θb_z_s_world_p as an Euler angle, or a quaternion equivalent thereto, is used.

[0218] According to the method of calculation represented by the respective block diagrams in FIG. 13 through FIG. 15, the estimated floor surface inclination angle value θslope_xy_s is calculated using a second order low-pass filter, and the first phase output value is calculated according to equation (17a), and the second phase output value is calculated according to equation (17b).
(The first phase output value)=(1−W1)(previous first phase output value)+W1(instantaneous provisional value)  17a
(The second phase output value)=(1−W2)(previous second phase output value)+W2(the first phase output value)  17b

[0219] A desired value θb_obj_xy of the base body tilt angle θb in the vehicle coordinate value is calculated on the basis of the estimated floor surface inclination angle value θslope_xy_s calculated as is explained above, a radius Rw_xy of the imaginary wheel 62_—xy, and a distance hgc_xy between the rotational shaft 62a_—xy of the imaginary wheel 62_—xy and the vehicle center-of-gravity (mass point) 60_—xy (stored in the memory). The desired base body tilt angle value θb_obj_xy is calculated so as to satisfy the condition that the X-coordinate value and the Y-coordinate value of the vehicle system center-of-gravity, respectively, and the X-coordinate value of the contact point between the imaginary wheel 62_—x and the floor surface and the Y-coordinate value of the contact point between the imaginary wheel 62_—y and the floor surface, respectively, coincide (refer to FIG. 16(c)).

[0220] The element for setting the desired value θb_obj_xy of the base body tilt angle as is explained above, and for controlling the base body tilt angle θb_s so as to coincide with the desired value θb_obj_xy corresponds to “a component which corrects the tilting of base body 9 at least in the second direction (Y-axis direction) towards the elevated side of the floor”.

[0221] Specifically, the desired base body tilt angle value θb_obj_xy is calculated by multiplying the estimated floor surface inclination angle value θslope_xy_s with Rw_xy/hgc_xy, according to the geometric relationship satisfying the above-mentioned conditions. The desired base body tilt angle value θb_obj_xy is input to the error calculator 70 as is explained above (refer to FIG. 17).

[0222] In place of the estimated value θslope_xy_s of the floor surface inclination angle θslope, the instantaneous provisional value θslope_xy_c may be output as the estimated floor surface inclination angle value θslope_xy_s.

[0223] The center-of-gravity velocity calculator 72 is input with the current value of the base body tilt angular velocity measured value θbdot_xy_s (θbdot_x_s and θbdot_y_s) calculated in STEP2, and also a previous value ωw_xy_cmd_p (ωw_x_cmd_p and ωw_y_cmd_p) of a imaginary wheel velocity command ωw_xy_cmd is input from a delay element 84. Thereafter, the center-of-gravity velocity calculator 72 calculates from these input values an estimated center-of-gravity velocity value Vb_xy_s (Vb_x_s and Vb_y_s), according to a predetermined arithmetic expression on the basis of the above-mentioned inverted pendulum model.

[0224] Specifically, the center-of-gravity velocity calculator 72 calculates Vb_x_s and Vb_y_s, respectively, from expressions 05x, 05y below.
Vb_—x_—s=Rw_—x·ωw_—x_—cmd_—p+h_—x·θbdot_—x_—s  05x
Vb_—y_—s=Rw_—y·ωw_—y_—cmd_—p+h_—y·θbdot_—y_—s  05y

[0225] In the expressions 05x, 05y, Rw_x, Rw_y are respective radius of the imaginary wheel 62_—x, 62_—y as is explained above, and these values are predetermined values set beforehand. Further, h_x, h_y are respective height of the mass point 60_—x, 60_—y of the inverted pendulum model. In the present embodiment, the height of the vehicle system center-of-gravity point is to be maintained substantially constant. Therefore, as the values of h_x, h_y, predetermined values set before hand are used. Supplementally, height h_x, h_y are included in the constant parameter, the value of which is set in STEP 6 or STEP8.

[0226] The first term in the right side of the equation 05x is a moving velocity of the imaginary wheel 62_—x in the X-axis direction corresponding to the previous value ωw_x_cmd_p of the velocity command of the imaginary wheel 62_—x, and this moving velocity corresponds to the current value of the actual moving velocity of the wheel assembly 5 in the X-axis direction. Further, the second term in the right side of the equation 05x corresponds to the current value of the moving velocity of the vehicle system center-of-gravity point in the X-axis direction arising from the base body 9 tilting at the tilt angular velocity of θbdot_x_s about the Y-axis direction (relative moving velocity with respect to the wheel assembly 5). The same applies to equation 05y.

[0227] Here, a pair of the measured value (current value) of the rotational angular velocity of the electric motors 31R, 31L, respectively, that are measured on the basis of the outputs from the rotary encoders 56R, 56L, may be converted to a pair of the rotational angular velocity of the imaginary wheels 62_—x, 62_—y, respectively, and these rotational angular velocities may be used instead of ωw_x_cmd_p, ωw_y_cmd_p in equations 05x, 05y. However, in eliminating the influence of the noise included in the measured value of the rotational angular velocity, it is advantageous to use ωw_x_cmd_p, ωw_y_cmd_p which are desired values.

[0228] Subsequently, the control unit 50 carries out the processing by the center-of-gravity velocity restrictor 76 and the processing by the gain adjustor 78. In this case, the center-of-gravity velocity restrictor 76 and the gain adjustor 78 respectively receive the estimated center-of-gravity velocity values Vb_xy_s (Vb_x_s and Vb_y_s) calculated by the center-of-gravity velocity calculator 72 as described above.

[0229] Further, the gain adjustor 78 determines the gain adjustment parameters Kr_xy (Kr_x and Kr_y) on the basis of the input estimated center-of-gravity velocity values Vb_xy_s (Vb_x_s and Vb_y_s).

[0230] The processing by the gain adjustor 78 will be described below with reference to FIG. 18 and FIG. 19.

[0231] As illustrated in FIG. 18, the gain adjustor 78 supplies the input estimated center-of-gravity velocity values Vb_x_s and Vb_y_s to a limiting processor 86. The limiting processor 86 adds, as appropriate, restrictions based on the permissible ranges of the rotational angular velocities of the electric motors 31R and 31L to the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s, thereby generating output values Vw_x_lim1 and Vw_y_lim1. The output value Vw_x_lim1 means a value obtained after limiting the moving velocity Vw_x of the imaginary wheel 62_—x in the X-axis direction and the output value Vw_y_lim1 means a value obtained after limiting the moving velocity Vw_y of the imaginary wheel 62_—y in the Y-axis direction.

[0232] The processing by the limiting processor 86 will be described in further detail with reference to FIG. 19. The parenthesized reference characters in FIG. 19 denote the processing by a limiting processor 104 of the center-of-gravity velocity restrictor 76, which will be discussed later, and may be ignored in the description related to the processing by the limiting processor 86.

[0233] The limiting processor 86 first supplies the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s to processors 86a_—x and 86a_—y, respectively. The processor 86a_—x divides Vb_x_s by the radius Rw_x of the imaginary wheel 62_—x to calculate the rotational angular velocity ωw_x_s of the imaginary wheel 62_—x in the case where it is assumed that the moving velocity of the imaginary wheel 62_—x in the X-axis direction coincides with Vb_x_s. Similarly, the processor 86a_—y calculates the rotational angular velocity ωw_y_s of the imaginary wheel 62_—y (=Vb_y_s/Rw_y) in the case where it is assumed that the moving velocity of the imaginary wheel 62_—y in the Y-axis direction coincides with Vb_y_s.

[0234] Subsequently, the limiting processor 86 converts the pair of ωw_x_s and ωw_y_s into a pair of the rotational angular velocity ω_R_s of the electric motor 31R and the rotational angular velocity ω_L_s of the electric motor 31L by an XY-RL converter 86b.

[0235] According to the present embodiment, the conversion is implemented by solving a simultaneous equation obtained by replacing ωw_x, ωw_y, ω_R and ω_L of the aforesaid expressions 01a and 01b by ωw_x_s, ωw_y_s, ω_R_s and ω_L_s, respectively, taking the ω_R_s and ω_L_s as unknowns.

[0236] Subsequently, the limit processor 86 supplies the output values ω_R_s and ω_L_s of the XY-RL converter 86b to limiters 86c_R and 86c_L, respectively. At this time, the limiter 86c_R directly outputs the ω_R_s as an output value ω_R_lim1 if the ω_R_s falls within the permissible range for the right motor, which has an upper limit value (>0) and a lower limit value (<0) of predetermined values set beforehand. If the ω_R_s deviates from the permissible range for the right motor, then the limiter 86c_R outputs, as the output value ω_R_lim1, a boundary value of the upper limit value or the lower limit value of the permissible range for the right motor whichever is closer to the ω_R_s. Thus, the output value ω_R_lim1 of the limiter 86c_R is limited to a value within the permissible range for the right motor.

[0237] Similarly, the limiter 86c_L directly outputs the ω_L_s as an output value ω_L_lim1 if the ω_L_s falls within the permissible range for the left motor, which has an upper limit value (>0) and a lower limit value (<0) of predetermined values set beforehand. If the ω_L_s deviates from the permissible range for the left motor, then the limiter 86c_L outputs, as the output value ω_L_lim1, a boundary value of the upper limit value or the lower limit value of the permissible range for the left motor whichever is closer to the ω_L_s. Thus, the output value ω_L_lim1 of the limiter 86c_L is limited to a value within the permissible range for the left motor.

[0238] The permissible range for the right motor described above is a permissible range which has been set so as to prevent the rotational angular velocity (absolute value) of the right electric motor 31R from becoming excessively high thereby to prevent the maximum value of the torque that can be output by the electric motor 31R from decreasing. This applies also to the permissible range for the left motor.

[0239] Subsequently, the limit processor 86 converts the pair of the output values ω_R_lim1 and ω_L_lim1 of the limiters 86c_R and 86c_L, respectively, into a pair of the rotational angular velocities ωw_x_lim1 and ωw_y_lim1 of the imaginary wheels 62_—x and 62_—y, respectively, by an RL-XY converter 86d.

[0240] The conversion is the processing of the inverse conversion of the processing of the conversion by the aforesaid XY-RL converter 86b. This processing is implemented by solving a simultaneous equation obtained by replacing ωw_x, ωw_y, ω_R and ω_L of the aforesaid expressions 01a and 01b by ωw_x_lim1, ωw_y_lim1, ω_R_lim1 and ω_L_lim1, respectively, taking the ωw_x_lim1 and ωw_y_lim1 as unknowns.

[0241] Subsequently, the limit processor 86 supplies the output values ωw_x_lim1 and ωw_y_lim1 of the RL-XY converter 86d to processors 86e_x and 86e_y, respectively. The processor 86e_x multiplies ωw_x_lim1 by the radius Rw_x of the imaginary wheel 62_—x to convert ωw_x_lim1 into the moving velocity Vw_x_lim1 of the imaginary wheel 62_—x. In the same manner, the processor 86e_y converts ωw_y_lim1 into the moving velocity Vw_y_lim1 of the imaginary wheel 62_—y (=ωw_y_lim1·Rw_y).

[0242] If it is assumed that the processing by the limiting processor 86 described above causes the moving velocity Vw_x of the imaginary wheel 62_—x in the X-axis direction and the moving velocity Vw_y of the imaginary wheel 62_—y in the Y-axis direction to agree with the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s, respectively (in other words, if it is assumed that the moving velocity of the wheel assembly 5 in the X-axis direction and the moving velocity in the Y-axis direction are set to agree with Vb_x_s and Vb_y_s, respectively), then the pair of output values Vw_x_lim1 and Vw_y_lim1 coinciding with Vb_x_s and Vb_y_s, respectively, is output from the limiting processor 86 if the rotational angular velocities ω_R_s and ω_L_s of the electric motors 31R and 31L, respectively, which are required for achieving the moving velocities, both fall within permissible ranges.

[0243] Meanwhile, if both or one of the rotational angular velocities ω_R_s and ω_L_s of the electric motors 31R and 31L, respectively, deviates from the permissible range or ranges, then both or one of the rotational angular velocities is forcibly limited to be within the permissible range, and a pair of the moving velocities in the X-axis direction and the Y-axis direction Vw_x_lim1 and Vw_y_lim1 corresponding to a pair of the limited rotational angular velocities ω_R_lim1 and ω_L_lim1 of the electric motors 31R and 31L, respectively, is output from the limiting processor 86.

[0244] Hence, the limiting processor 86 generates a pair of output values Vw_x_lim1 and Vw_y_lim1 such that the output values Vw_x_lim1 and Vw_y_lim1 agree with Vb_x_s and Vb_y_s, respectively, as much as possible under an essential required condition that the rotational angular velocities of the electric motors 31R and 31L corresponding to the pair of the output values Vw_x_lim1 and Vw_y_lim1, respectively, do not deviate from the permissible ranges.

[0245] Referring back to the description of FIG. 18, the gain adjustor 78 then carries out the processing by calculators 88_—x and 88_—y. The calculator 88_—x receives the estimated center-of-gravity velocity value in the X-axis direction Vb_x_s and the output value Vw_x_lim1 of the limiting processor 86. Then, the calculator 88_—x calculates a value Vover_x obtained by subtracting Vb_x_s from Vw_x_lim1 and outputs the value Vover_x. Further, the calculator 88_—y receives the estimated center-of-gravity velocity value in the Y-axis direction Vb_y_s and the output value Vw_y_lim1 of the limiting processor 86. Then, the calculator 88_—y calculates a value Vover_y obtained by subtracting Vb_y_s from Vw_y_lim1 and outputs the value Vover_y.

[0246] In this case, if the output values Vw_x_lim1 and Vw_y_lim1 are not forcibly restricted by the limiting processor 86, then Vw_x_lim1=Vb_x_s and Vw_y_lim1=Vb_y_s. Therefore, the output values Vover_x and Vover_y of the calculators 88_—x and 88_—y, respectively, will be both zero.

[0247] Meanwhile, if the output values Vw_x_lim1 and Vw_y_lim1 of the limiting processor 86 are generated by forcibly restricting the input values Vb_x_s and Vb_y_s, then a corrected amount from Vb_x_s of Vw_x_lim1 (=Vw_x_lim1−Vb_x_s) and a corrected amount from Vb_y_s of Vw_y_lim1 (=Vw_y_lim1−Vb_y_s) will be output from the calculators 88_—x and 88_—y, respectively.

[0248] Subsequently, the gain adjustor 78 passes the output value Vover_x of the calculator 88_—x through processors 90_—x and 92_—x in this order thereby to determine the gain adjustment parameter Kr_x. Further, the gain adjustor 78 passes the output value Vover_y of the calculator 88_—y through processors 90_—y and 92_—y in this order thereby to determine the gain adjustment parameter Kr_y. The gain adjustment parameters Kr_x and Kr_y both take values within the range from 0 to 1.

[0249] The processor 90_—x calculates and outputs the absolute value of the input Vover_x. Further, the processor 92_—x generates Kr_x such that the output value Kr_x monotonously increases relative to an input value |Vover_x| and has a saturation characteristic. The saturation characteristic is a characteristic in which a change amount of an output value relative to an increase in an input value becomes zero or approaches to zero when the input value increases to a certain level.

[0250] In this case, according to the present embodiment, if the input value |Vover_x| is a preset, predetermined value or less, then the processor 92_—x outputs, as Kr_x, a value obtained by multiplying the input value |Vover_x| by a proportionality coefficient of a predetermined value. Further, if the input value |Vover_x| is larger than the predetermined value, then the processor 92_—x outputs 1 as Kr_x. Incidentally, the proportionality coefficient is set such that the product of |Vover_x| and the proportionality coefficient becomes 1 when |Vover_x| agrees with a predetermined value.

[0251] The processing by processors 90_—y and 92_—y is the same as that carried out by the processors 90_—x and 92_—x, respectively, described above.

[0252] If the output values Vw_x_lim1 and Vw_y_lim1 in the limiting processor 86 are not forcibly restricted by the processing carried out by the gain adjustor 78 described above, that is, if the rotational angular velocities of the electric motors 31R and 31L fall within the permissible ranges even when the electric motors 31R and 31L are operated such that the moving velocities Vw_x and Vw_y of the wheel assembly 5 in the X-axis direction and the Y-axis direction, respectively, agree with the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s, respectively, then the gain adjustment parameters Kr_x and Kr_y are both determined to be zero. Therefore, normally, Kr_x=Kr_y=0.

[0253] Meanwhile, if the output values Vw_x_lim1 and Vw_y_lim1 of the limiting processor 86 are generated by forcibly restricting the input values Vb_x_s and Vb_y_s, that is, if the rotational angular velocity of one of the electric motors 31R and 31L deviates from the permissible range (if the absolute value of one of the rotational angular velocities becomes excessively high) when the electric motors 31R and 31L are operated such that the moving velocities Vw_x and Vw_y of the wheel assembly 5 in the X-axis direction and the Y-axis direction, respectively, agree with the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s, then the values of the gain adjustment parameters Kr_x and Kr_y are determined on the basis of the absolute values of the aforesaid corrected amounts Vover_x and Vover_y, respectively. In this case, Kr_x is determined to be a larger value as the absolute value of the corrected amount Vover_x increases, the upper limit value thereof being 1. The same applies to Kr_y.

[0254] The center-of-gravity velocity restrictor 76 determines the desired center-of-gravity velocities for control Vb_xy_mdfd (Vb_x_mdfd and Vb_y_mdfd), by executing the process shown in the block diagram of FIG. 20, using the input estimated center-of-gravity velocity values Vb_xy_s (Vb_x_s and Vb_y_s).

[0255] To be more specific, the center-of-gravity velocity restrictor 76 first carries out the processing by stead-state error calculators 94_—x and 94_—y.

[0256] In this case, the stead-state error calculator 94_—x receives the estimated center-of-gravity velocity value in the X-axis direction Vb_x_s and also receives the previous value Vb_x_mdfd_p of the desired center-of-gravity velocity for control in the X-axis direction Vb_x_mdfd through the intermediary of a delay element 96_—x. Then, in the stead-state error calculator 94_—x, first, the input Vb_x_s is supplied to a proportional-differential compensation component (PD compensation component) 94a_—x. The proportional-differential compensation component 94a_—x is a compensation component whose transfer function is denoted by 1+Kd·S, and adds the input Vb_x_s and the value obtained by multiplying the differential value thereof (temporal change rate) by a coefficient Kd of a predetermined value, and outputs the value resulting from the addition.

[0257] Subsequently, the stead-state error calculator 94_—x calculates, by a calculator 94b_—x, the value obtained by subtracting the input Vb_x_mdfd_p from the output value of the proportional-differential compensation component 94a_—x, then supplies the output value of the calculator 94b_—x to a low-pass filter 94c_x having a phase compensation function. The low-pass filter 94c_x is a filter whose transfer function is denoted by (1+T2·S)/(1+T1·S). Then, the stead-state error calculator 94_—x outputs the output value Vb_x_prd of the low-pass filter 94c_x.

[0258] Further, the stead-state error calculator 94_—y receives the estimated center-of-gravity velocity value in the Y-axis direction Vb_y_s and also receives the previous value Vb_y_mdfd_p of the desired center-of-gravity velocity for control in the Y-axis direction Vb_y_mdfd through the intermediary of a delay element 96_—y.

[0259] Then, as with the stead-state error calculator 94_—x described above, the stead-state error calculator 94_—y carries out the processing by a proportional-differential compensation component 94a_y, a calculator 94b_y, and a low-pass filter 94c_y in order and outputs an output value Vb_y_prd of the low-pass filter 94c_y.

[0260] Here, the output value Vb_x_prd of the stead-state error calculator 94_—x has a meaning as a stead-state error of a future estimated center-of-gravity velocity value in the X-axis direction presumed from a current motion state of the vehicle system center-of-gravity point observed from the Y-axis direction (in other words, the motion state of the mass point 60_—x of the inverted pendulum model observed from the Y-axis direction) relative to the desired center-of-gravity velocity for control Vb_x_mdfd of an expected convergence value. Similarly, the output value Vb_y_prd of the stead-state error calculator 94_—y has a meaning as a stead-state error of a future estimated center-of-gravity velocity value in the Y-axis direction presumed from a current motion state of the vehicle system center-of-gravity point observed from the X-axis direction (in other words, the motion state of the mass point 60_—y of the inverted pendulum model observed from the X-axis direction) relative to the desired center-of-gravity velocity for control Vb_y_mdfd of an expected convergence value. Hereinafter, the output values Vb_x_prd and Vb_y_prd of the stead-state error calculators 94_—x and 94_—y, respectively, will be referred to as the predicted center-of-gravity velocity stead-state error values.

[0261] After executing the process of the stead-state error calculators 94_—x and 94_—y as explained above, the center-of-gravity velocity restrictor 76 inputs the predicted center-of-gravity velocity stead-state error values Vb_x_prd and Vb_y_prd, respectively, to a limiting processor 100. The processing by the limiting processor 100 is the same as the processing by the limiting processor 86 of the gain adjustor 78 described above. In this case, as indicated by the parenthesized reference characters in FIG. 19, only the input values and the output values of the individual processing sections of the limiting processor 100 are different from those of the limiting processor 86.

[0262] To be more specific, in the limiting processor 100, rotational angular velocities ωw_x_t and ωw_y_t of the imaginary wheels 62_—x and 62_—y in the case where it is assumed that the moving velocities Vw_x and Vw_y of the imaginary wheels 62_—x and 62_—y, respectively, coincide with Vb_x_prd and Vb_y_prd, respectively, are calculated by the processors 86a_x and 86a_y, respectively. Then, the pair of the rotational angular velocities ωw_x_t and ωw_y_t is converted into the pair of the rotational angular velocities ω_R_t and ω_L_t of the electric motors 31R and 31L by the XY-RL converter 86b.

[0263] Further, these rotational angular velocities ω_R_t and ωL_t are limited to values within the permissible range for the right motor and the permissible range for the left motor, respectively, by limiters 86c_R and 86c_L. Then, the values ω_R_lim2 and ω_L_lim2, which have been subjected to the limitation processing, are converted by the RL-XY converter 86d into the rotational angular velocitiesωw_x_lim2 and ωw_y_lim2 of the imaginary wheels 62_—x and 62_—y.

[0264] Subsequently, the moving velocities Vw_x_lim2 and Vw_y_lim2 of the imaginary wheels 62_—x and 62_—y corresponding to the rotational angular velocities ωw_x_lim2 and ωw_y_lim2 are calculated by the processors 86e_x and 86e_y, respectively, and these moving velocities Vw_x_lim2 and Vw_y_lim2 are output from the limiting processor 100.

[0265] By carrying out the processing by the limiting processor 100 described above, the limiting processor 100 generates a pair of output values Vw_x_lim2 and Vw_y_lim2 such that the output values Vw_x_lim2 and Vw_y_lim2 agree with Vb_x_t and Vb_y_t, respectively, as much as possible under an essential required condition that the rotational angular velocities of the electric motors 31R and 31L corresponding to the pair of the output values Vw_x_lim2 and Vw_y_lim2, respectively, do not deviate from the permissible ranges, as with the limiting processor 86.

[0266] Incidentally, the permissible ranges for the right motor and left motor in the limiting processor 100 do not have to be the same as the permissible ranges in the limiting processor 86, and may be set to be permissible ranges that are different from each other.

[0267] Referring back to the description of FIG. 20, the center-of-gravity velocity restrictor 76 then carries out the processing by calculators 102_—x and 102_—y to calculate the desired center-of-gravity velocities for control Vb_x_mdfd and Vb_y_mdfd, respectively. In this case, the calculator 102_—x calculates a value, which is obtained by subtracting the predicted center-of-gravity velocity stead-state error value in the X-axis direction Vb_x_prd from the output value Vw_x_lim2 of the limiting processor 100, as the desired center-of-gravity velocity for control in the X-axis direction Vb_x_mdfd. Similarly, the calculator 102_—y calculates a value, which is obtained by subtracting the predicted center-of-gravity velocity stead-state error value in the Y-axis direction Vb_y_prd from the output value Vw_y_lim2 of the limiting processor 100, as the desired center-of-gravity velocity for control in the Y-axis direction Vb_y_mdfd.

[0268] Regarding the desired center-of-gravity velocities for control Vb_x_mdfd and Vb_y_mdfd determined as described above, in the case where the output values Vw_x_lim2 and Vw_y_lim2 are not forcibly restricted by the limiting processor 100, that is, in the case where the rotational angular velocities of the electric motors 31R and 31L fall within the permissible ranges even when the electric motors 31R and 31L are operated such that the moving velocities of the wheel assembly 5 in the X-axis direction and the Y-axis direction, respectively, agree with the predicted center-of-gravity velocity stead-state error values Vb_x_prd and Vb_y_prd, respectively, the desired center-of-gravity velocities for control Vb_x_mdfd and Vb_y_mdfd are both determined to be “0”. Therefore, normally, Vb_x_mdfd=Vb_y_mdfd=0.

[0269] Meanwhile, if the output values Vw_x_lim2 and Vw_y_lim2 of the limiting processor 100 are generated by forcibly restricting the input values Vb_x_t and Vb_y_t, that is, if the rotational angular velocity of one of the electric motors 31R and 31L deviates from the permissible range (if the absolute value of one of the rotational angular velocities becomes excessively high) when the electric motors 31R and 31L are operated such that the moving velocities of the wheel assembly 5 in the X-axis direction and the Y-axis direction, respectively, agree with the predicted center-of-gravity velocity stead-state error values Vb_x_prd and Vb_y_prd, then for the X-axis direction, a correction amount from the input value Vb_x_prd of the output value Vw_x_lim2 of the limiting processor 100 (=Vw_x_lim2−Vb_x_prd) is determined as the desired center-of-gravity velocity for control Vb_x_mdfd in the X-axis direction.

[0270] Further, regarding the Y-axis direction, a correction amount from the input value Vb_y_prd of the output value Vw_y_lim2 of the limiting processor 100 (=Vw_y_lim2−Vb_y_prd) is determined as the desired center-of-gravity velocity for control Vb_y_mdfd in the Y-axis direction.

[0271] In this case, for example with respect to the velocity in the X-axis direction, the desired center-of-gravity velocity for control Vb_x_mdfd becomes the velocity in opposite direction from the predicted center-of-gravity velocity stead-state error value Vb_x_prd in the X-axis direction which is output from the stead-state error calculator 94_—x. The same applies to the velocity in the Y-axis direction.

[0272] The above has described the processing by the center-of-gravity velocity restrictor 76.

[0273] Returning to the description of FIG. 17, after carrying out the processing by the center-of-gravity velocity restrictor 76, the gain adjustor 78, and the error calculator 70 as described above, the control unit carries out the processing by the posture control calculator 80.

[0274] The processing by the posture control calculator 80 will be described below by referring to FIG. 21. Incidentally, the unparenthesized reference characters in FIG. 21 are the reference characters related to the processing for determining the aforesaid imaginary wheel rotational angular velocity command ωw_x_cmd, which is the desired value of the rotational angular velocity of the imaginary wheel 62_—x circumrotating in the X-axis direction. The parenthesized reference characters are the reference characters related to the processing for determining the aforesaid imaginary wheel rotational angular velocity command ωw_y_cmd, which is the desired value of the rotational angular velocity of the imaginary wheel 62_—y circumrotating in the Y-axis direction.

[0275] The posture control calculator 80 receives the base body tilt angle error measured value θbe_xy_s calculated by the error calculator 70, the base body tilt angular velocity measured values θbdot_xy_s calculated in the aforesaid STEP2, the estimated center-of-gravity velocity values Vb_xy_s calculated by the center-of-gravity velocity calculator 72, the desired center-of-gravity velocities for control Vb_xy_mdfd calculated by the center-of-gravity velocity restrictor 76, and the gain adjustment parameters Kr_xy calculated by the gain adjustor 78.

[0276] Then, the posture control calculator 80 first calculates the imaginary wheel rotational angular acceleration commands ωwdot_xy_cmd according to the following expressions 07x and 07y by using the above received values.
ωwdot_—x_—cmd=K1_—x·θbe_—x_—s+K2x_—θbdot_—x_—s+K3_—x·(Vb_—x_—s−Vb_—x_—mdfd)  Expression 07x
ωwdot_—y_—cmd=K1_—y·θbe_—y_—s+K2_—y·θbdot_—y_—s+K3_—y·(Vb_—y_—s−Vb_—y_—mdfd)  Expression 07y

[0277] Hence, according to the present embodiment, the imaginary wheel rotational angular acceleration command ωwdot_x_cmd denoting the manipulated variable (control input) for controlling the motion of the mass point 60_—x of the inverted pendulum model observed from the Y-axis direction (i.e., the motion of the vehicle system center-of-gravity point observed from the Y-axis direction), and the imaginary wheel rotational angular acceleration command ωwdot_y_cmd denoting the manipulated variable (control input) for controlling the motion of the mass point 60_—y of the inverted pendulum model observed from the X-axis direction (i.e., the motion of the vehicle system center-of-gravity point observed from the X-axis direction) are determined by adding up three manipulated variable components (the three terms of the right side of each of expressions 07x and 07y).

[0278] In this case, the gain coefficients K1_—x, K2_—x, and K3_—x related to the manipulated variable components in expression 07x are variably set on the basis of the gain adjustment parameter Kr_x, while the gain coefficients K1_—y, K2_—y, and K3_—y related to the manipulated variable components in expression 07y are variably set on the basis of the gain adjustment parameter Kr_y. Hereinafter, the gain coefficients K1_—x, K2_—x, and K3_—x in expression 07x may be referred to as the first gain coefficient K1_—x, the second gain coefficient K2_—x, and the third gain coefficient K3_—x, respectively. The same applies to the gain coefficients K1_—y, K2_—y, and K3_—y in expression 07y.

[0279] An i-th gain coefficient Ki_x (i=1, 2, or 3) in expression 07x and an i-th gain coefficient Ki_y (i=1, 2, or 3) in expression 07y are determined on the basis of the gain adjustment parameters Kr_x and Kr_y according to expressions 09x and 09y given below, as indicated by the note in FIG. 21.
Ki_—x=(1−Kr_—x)·Ki_—a_—x+Kr_—x·Ki_—b_—x  Expression 09x
Ki_—y=(1−Kr_—y)·Ki_—a_—y+Kr_—y·Ki_—b_—y  Expression 09y

• • [0280] (i=1, 2, 3)

[0281] Here, Ki_a_x and Ki_b_x in expression 09x denote constant values set beforehand as the gain coefficient values on a minimum end (an end close to zero) of the i-th gain coefficient Ki_x and as the gain coefficient value on a maximum end (an end away from zero), respectively. The same applies to Ki_a_y and Ki_b_y in expression 09y.

[0282] Thus, each i-th gain coefficient Ki_x (i=1, 2 or 3) used in the calculation of expression 07x is determined as a weighed mean value of the constant values Ki_a_x and Ki_b_x corresponding thereto. Further, in this case, the weight applied to each of Ki_a_x and Ki_b_x is changed according to the gain adjustment parameter Kr_x. Therefore, if Kr_x=0, then Ki_x=Ki_a_x, and if Kr_x=1, then Ki_x=Ki_b_x. As Kr_x approaches 1 from 0, the i-th gain coefficient Ki_x approaches Ki_b_x from Ki_a_x.

[0283] Similarly, each i-th gain coefficient Ki_y (i=1, 2 or 3) used in the calculation of expression 07y is determined as a weighed mean value of the constant values Ki_a_y and Ki_b_y corresponding thereto. Further, in this case, the weight applied to each of Ki_a_y and Ki_b_y is changed according to the gain adjustment parameter Kr_y. Therefore, as with the case of Ki_x, as the value of Kr_y changes from 0 to 1, the value of the i-th gain coefficient Ki_y changes between Ki_a_y and Ki_b_y.

[0284] As is explained above, Kr_x and Kr_y are normally (to be more specific, in the case where the output values Vw_x_lim1 and Vw_y_lim1 are not forcibly restricted at the limit processor 86 of the gain adjustor 78), “0”. Therefore, the i-th gain coefficient Ki_x and Ki_y (i=1, 2 or 3) normally become Ki_x=Ki_a_x and Ki_y=Ki_a_y, respectively.

[0285] Supplementally, the aforesaid constant values Ki_a_x, Ki_b_x and Ki_a_y, Ki_b_y (i=1, 2 or 3) are included in the constant parameters, the values of which are set in the aforesaid STEP6 or STEP8.

[0286] The posture control calculator 80 uses the first to the third gain coefficients K1_—x, K2_—x, and K3_—x determined as described above to perform the calculation of the above expression 07x, thereby calculating the imaginary wheel rotational angular acceleration command ωwdot_x_cmd related to the imaginary wheel 62_—x circumrotating in the X-axis direction.

[0287] More detailedly, referring to FIG. 21, the posture control calculator 80 calculates a manipulated variable component u1_—x obtained by multiplying the base body tilt angle error measured value θbe_x_s by the first gain coefficient K1_—x and a manipulated variable component u2_—x obtained by multiplying the base body tilt angular velocity measured value θbdot_x_s by the second gain coefficient K2_—x by processors 80a and 80b, respectively. Further, the posture control calculator 80 calculates the difference between the estimated center-of-gravity velocity value Vb_x_s and the desired center-of-gravity velocity for control Vb_x_mdfd (=Vb_x_s−Vb_x_mdfd) by a calculator 80d, and calculates, by a processor 80c, a manipulated variable component u3_—x obtained by multiplying the calculated difference by the third gain coefficient K3_—x. Then, the posture control calculator 80 adds up these manipulated variable components u1_—x, u2_—x, and u3_—x by a calculator 80e so as to calculate the imaginary wheel rotational angular acceleration command ωwdot_x_cmd.

[0288] Similarly, the posture control calculator 80 carries out the calculation of the above expression 07y by using the first to the third gain coefficients K1_—y, K2_—y, and K3_—y determined as described above, thereby calculating the imaginary wheel rotational angular acceleration command ωwdot_y_cmd related to the imaginary wheel 62_—y circumrotating in the Y-axis direction.

[0289] In this case, the posture control calculator 80 calculates a manipulated variable component u1_—y obtained by multiplying the base body tilt angle error measured value θbe_y_s by the first gain coefficient K1_—y and a manipulated variable component u2_—y obtained by multiplying the base body tilt angular velocity measured value θbdot_y_s by the second gain coefficient K2_—y by the processors 80a and 80b, respectively. Further, the posture control calculator 80 calculates the difference between the estimated center-of-gravity velocity value Vb_y_s and the desired center-of-gravity velocity for control Vb_y_mdfd (=Vb_y_s−Vb_y_mdfd) by the calculator 80d, and calculates, by the processor 80c, a manipulated variable component u3_—y obtained by multiplying the calculated difference by the third gain coefficient K3_—y. Then, the posture control calculator 80 adds up these manipulated variable components u1_—y, u2_—y, and u3_—y by the calculator 80e so as to calculate the imaginary wheel rotational angular acceleration command ωwdot_y_cmd.

[0290] Here, the first term (=the first manipulated variable component u1_—x) and the second term (=the second manipulated variable component u2_—x) of the right side of expression 07x mean the feedback manipulated variable components for converging the base body tilt angle error measured value θbe_x_s in the direction about the X-axis to zero (converging the base body tilt angle measured value θb_x_s to the desired value θb_x_obj) by the PD law (proportional-differential law) serving as the feedback control law.

[0291] Further, the third term (=the third manipulated variable component u3_—x) of the right side of expression 07x means a feedback manipulated variable component for converging the difference between the estimated center-of-gravity velocity value Vb_x_s and the desired center-of-gravity velocity for control Vb_x_mdfd to zero (converging Vb_x_s to Vb_x_mdfd) by a proportional law serving as the feedback control law.

[0292] The same applies to the first to the third terms (the first to the third manipulated variable components u1_—y, u2_—y, and u3_—y) of the right side of expression 07y.

[0293] As is explained above, normally (to be more specific, in the case where the output values V_x_lim2 and V_y_lim2 are not forcibly limited at the limit processor 100 of the center-of-gravity velocity restrictor 76), the desired center-of-gravity velocities for control Vb_x_mdfd and Vb_y_mdfd are “0”. And, in the normal case where Vb_x_mdfd=Vb_y_mdfd=0, the third manipulated variable components u3_—x and u3_—y, respectively, coincide with the value obtained by multiplying the third gain coefficients K3_—x and K3_—y with the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s, respectively.

[0294] After calculating the imaginary wheel rotational angular acceleration commands ωwdot_x_cmd and ωwdot_y_cmd as described above, the posture control calculator 80 integrates these ωwdot_x_cmd and ωwdot_y_cmd by an integrator 80f thereby to determine the aforesaid imaginary wheel rotational velocity commands ωw_x_cmd and ωw_y_cmd.

[0295] The above has described the details of the processing by the posture control calculator 80.

[0296] Supplementally, the imaginary wheel rotational angular acceleration command ωwdot_x_cmd may alternatively be calculated by an expression wherein the third term of the right side of expression 07x is separated into the manipulated variable component based on Vb_x_s (=K3_—x·Vb_x_s) and the manipulated variable component based on Vb_x_mdfd (=−K3_—x·Vb_x_mdfd). Similarly, the imaginary wheel rotational angular acceleration command ωwdot_y_cmd may alternatively be calculated by an expression wherein the third term of the right side of expression 07y is separated into the manipulated variable component based on Vb_y_s (=K3_—y·Vb_y_s) and the manipulated variable component based on Vb_y_mdfd (=−K3_—y·Vb_y_mdfd).

[0297] Further, in the present embodiment, the rotational angular acceleration commands wdotw x_cmd and wdotw y_cmd of the imaginary wheels 62_—x and 62_—y have been used as the manipulated variables (control inputs) for controlling the behaviors of the vehicle system center-of-gravity point. However, for example, the drive torques of the imaginary wheels 62_—x and 62_—y or the translational forces obtained by dividing the driving torques by the radii Rw_x and Rw_y of the imaginary wheels 62_—x and 62_—y (i.e., the frictional forces between the imaginary wheels 62_—x, 62_—y and a floor surface) may be used as the manipulated variables.

[0298] Returning to the description of FIG. 17, the control unit 50 then supplies the imaginary wheel rotational velocity commands ωw_x_cmd and ωw_y_cmd determined as described above by the posture control calculator 80 to the motor command calculator 82, and carries out the processing by the motor command calculator 82 so as to determine a velocity command ω_R_cmd of the electric motor 31R and a velocity command ω_L_cmd of the electric motor 31L. The processing by the motor command calculator 82 is the same as the processing by the XY-RL converter 86b of the aforesaid limiting processor 86 (refer to FIG. 19).

[0299] To be more specific, the motor command calculator 82 determines the velocity commands ω_R_cmd and ω_L_cmd of the electric motors 31R and 31L by a simultaneous equation obtained by replacing ωw_x, ωw_y, ω_R and ω_L of the aforesaid expressions 01a and 01b by ωw_x_cmd, ωw_y_cmd, ω_R_cmd and ω_L_cmd, respectively, taking the ω_R_cmd and ω_L_cmd as unknowns.

[0300] Thus, the vehicle control arithmetic processing in the aforesaid STEP9 is completed.

[0301] (Operation of the Vehicle)

[0302] By the control arithmetic processing carried out by the control unit 50 as described above, the imaginary wheel rotational angular acceleration commands ωwdot_xy_cmd denoting the manipulated variables (control inputs) are determined such that, basically, in the case where the posture of the seat 3 and the base body 9 is maintained at a posture in which the aforesaid base body tilt angle error measured values θbe_x_s and θbe_y_s are both zero (hereinafter, this posture will be referred to as the basic posture) even in a motion mode which is either the boarding mode or the autonomous mode, the vehicle system center-of-gravity point becomes static. Thereafter, when the posture of the seat 3 and the base body 9 is tilted with respect to the basic posture, that is, when the horizontal direction position of the vehicle system center-of-gravity point (the vehicle-occupant overall center-of-gravity point or the vehicle-alone center-of-gravity point) is displaced from the state in which the same is positioned substantially above the ground contact surface of the wheel assembly 5, the imaginary wheel rotational angular acceleration commands ωwdot_xy_cmd are determined so as to restore the posture of the seat 3 and the base body 9 to the basic posture (so as to bring θbe_x_s and θbe_y_s close to “0”, or to maintain the same at “0”).

[0303] Then, the rotational angular velocities of the electric motors 31R and 31L, respectively, obtained by converting the imaginary wheel rotational angular velocity commands ωw_xy_cmd, which is obtained by integrating each component of ωwdot_xy_cmd, are determined as the velocity commands ω_R_cmd and ω_L_cmd of the electric motors 31R and 31L. Further, the rotational velocities of the electric motors 31R and 31L are controlled according to the velocity commands ω_R_cmd and ω_L_cmd. Thus, the moving velocities of the wheel assembly 5 in the X-axis direction and the Y-axis direction, respectively, are controlled so as to agree with the moving velocity of the imaginary wheel 62_—x corresponding to ωw_x_cmd and the moving velocity of the imaginary wheel 62_—y corresponding to ωw_y_cmd, respectively.

[0304] With this arrangement, if, for example, the actual base body tilt angle θb_x deviates from the desired value θb_x_obj in the direction about the Y-axis by leaning forward, then the wheel assembly 5 moves forward to eliminate the deviation (to converge θbe_x_s to zero). Similarly, if the actual θb_x deviates from the desired value θb_x_obj by leaning backward, then the wheel assembly 5 moves backward to eliminate the deviation (to converge θbe_x_s to zero).

[0305] Further, for example, if the actual base body tilt angle θb_y deviates from the desired value θb_y_obj in the direction about the X-axis by leaning rightward, then the wheel assembly 5 moves rightward to eliminate the deviation (to converge θbe_y_s to zero). Similarly, if the actual θb_y deviates from the desired value θb_y_obj by leaning leftward, then the wheel assembly 5 moves leftward to eliminate the deviation (to converge θbe_y_s to zero).

[0306] Further, if both the actual base body tilt angles θb_x and θb_y deviate from the desired values θb_x_obj and θb_y_obj, respectively, then the moving operation of the wheel assembly 5 in the fore-and-aft direction to eliminate the deviation of θb_x and the moving operation of the wheel assembly 5 in the lateral direction to eliminate the deviation of θb_y are combined, so that the wheel assembly 5 will move in a direction combining the X-axis direction and the Y-axis direction (a direction at an angle to both the X-axis direction and the Y-axis direction).

[0307] Thus, if the posture of the seat 3 and the base body 9 tilts from the basic posture, then the wheel assembly 5 moves toward the tilting side. Hence, if, for example, the occupant intentionally inclines his/her upper body together with the seat 3 and the base body 9 in the aforesaid boarding mode, then the wheel assembly 5 will move to the tilting side. In the present embodiment, in the case where the seat 3 and the base body 9 are tilted from the basic posture, the traveling direction of the vehicle system center-of-gravity point within the horizontal surface (traveling direction in the direction orthogonal to the Z-axis) and the traveling direction of the wheel assembly 5 does not necessarily coincide, from the reasons to be given later.

[0308] When the posture of the seat 3 and the base body 9 is held in a constant posture tilted from the basic posture (a posture in which the base body tilt angle error measured value θbe_xy_s becomes constant), during traveling of the wheel assembly 5 (during traveling of the overall vehicle 1), the moving velocity of the vehicle system center-of-gravity point (and consequently the moving velocity of the wheel assembly 5) converges to a moving velocity having a constant error with the desired center-of-gravity velocities for control Vb_xy_mdfd and also the error thereof depending on the base body tilt angle error measured value θbe_xy_s.

[0309] In the present embodiment, in the normal case where the traveling motion of the wheel assembly 5 is carried out in a moving velocity in which the rotational angular velocity of the electric motors 31R and 31L does not become too fast (to be more precise, in the case where the output values Vw_x_lim2 and Vw_y_lim2 are not forcibly restricted by the limit processor 100 of the center-of-gravity velocity restrictor 76), the desired center-of-gravity velocities for control Vb_x_mdfd and Vb_y_mdfd are both maintained at “0”. Thereafter, when the posture of the seat 3 and the base body 9 are held in a constant position tilted from the basic position, in the situation where Vb_x_mdfd and Vb_y_mdfd are maintained constant, the moving velocity of the vehicle system center-of-gravity point (and consequently the moving velocity of the wheel assembly 5) converges to the moving velocity having the magnitude and orientation depending on the base body tilt angle error measured value θbe_xy_s.

[0310] To give more detailed explanation on the operation mentioned above, in a stationary state in which both of the base body tilt angle error measured values θbe_x_s and θbe_y_s are held constant, the second manipulated variable components u2_—x and u2_—y becomes “0”. Therefore, the imaginary wheel rotational angular acceleration command ωwdot_x_cmd becomes a value in which the first manipulated variable component u1_—x is added with the third manipulated variable component u3_—x, and the imaginary wheel rotational angular acceleration command ωwdot_y_cmd becomes a value in which the first manipulated variable component u1_—y is added with the third manipulated variable component u3_—y.

[0311] Further, in the stationary state mentioned above, the imaginary wheel rotational angular acceleration commands ωwdot_x_cmd and ωwdot_y_cmd converge to values in which the moving velocity of the wheel assembly 5 may be kept constant. Consequently, the center-of-gravity velocities Vb_x and Vb_y converge to constant values.

[0312] In this case, the second term in the right side of the equation 07x (=u2_—x) becomes “0”, the first term on the right side (=u1_—x=K1_—x*θbe_x_s) becomes a constant value, and the ωwdot_x_cmd on the left side becomes a constant value, so that a converging value of the center-of-gravity velocity Vb_x in the X-axis direction (a converging value of the estimated center-of-gravity velocity value Vb_x_s, hereinafter referred to as a stationary-state converging velocity Vb_x_stb) depends on the base body tilt angle error measured value θbe_x_s about the Y-axis direction. More specifically, Vb_x_stb becomes Vb_x_stb=(−K1_—x*Δθbe_x_s+ωwdot_x_cmd)/K3_—x+Vb_x_mdfd, so that it becomes a function value which monotonously change with respect to θbe_x_s.

[0313] Similarly, the second term in the right side of the equation 07y (=u2_—y) becomes “0”, the first term on the right side (=u1_—y=K1_—y*θbe_y_s) becomes a constant value, and ωwdot_y_cmd in the left side becomes a constant value, so that a converging value of the center-of-gravity velocity Vb_y in the Y-axis direction (a converging value of the estimated center-of-gravity velocity value Vb_y_s, hereinafter referred to as the stationary-state converging velocity Vb_y_stb) depends on the base body tilt angle error measured value θbe_y_s about the X-axis direction.

[0314] More specifically, Vb_x_stb becomes Vb_y_stb=(−K1_—y*Δθbe_y_s+ωwdot_y_cmd)/K3_—y+Vb_y_mdfd, so that it becomes a function value which monotonously change with respect to θbe_y_s.

[0315] As seen from above, when the posture of the seat 3 and the base body 9 are kept in a constant posture tilted from the base body in the situation where the Vb_x_mdfd and Vb_y_mdfd are kept constant, the moving velocity of the vehicle system center-of-gravity point (and consequently the moving velocity of the wheel assembly 5) converges to a moving velocity with magnitude and orientation depending on the base body tilt angle error measured value θbe_xy_s.

[0316] Further, for example, in the situation where the tilt amount of the base body 9 and the seat 3 from the basic posture (the base body tilt angle error measured values θbe_x_s and θbe_y_s) becomes relatively large, and the moving velocities of the wheel assembly 5 in one or both of the X-axis direction and the Y-axis direction in the case where the tilt amount is maintained stable (these moving velocities correspond to the predicted center-of-gravity velocity stead-state error values Vb_x_prd and Vb_y_prd, respectively, shown in FIG. 20) becomes excessive moving velocity so as to make one or both of the rotational angular velocities of the electric motors 31R and 31L deviate from the allowable range at the limiting processor 100, then the velocities in the direction opposite to the moving velocity of the wheel assembly 5 (more specifically, Vw_x_lim2-Vb_x_prd and Vw_y_lim2−Vb_y_prd) are determined as the desired center-of-gravity velocities for control Vb_x_mdfd and Vb_y_mdfd. Thereafter, of the manipulated variable components constituting the control input, the third manipulated variable components u3_—x and u3_—y are determined so as to converge the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s to the desired center-of-gravity velocities for control Vb_x_mdfd and Vb_y_mdfd, respectively. By doing so, velocity increasing of the wheel assembly 5 is suppressed, and consequently one or both of the rotational angular velocities of the electric motors 31R and 31L are prevented from being too fast.

[0317] Further, in the situation where one or both of the estimated center-of-gravity velocity values Vb_x_s and Vb_y_s become large, and consequently one or both of the moving velocities of the wheel assembly 5 in the X-axis direction and the Y-axis direction become excessive moving velocity so as to make one or both of the rotational angular velocities of the electric motors 31R and 31L deviate from the allowable range at the limiting processor 86, then in the gain adjustor 78, one or both of the gain adjustment parameters Kr_x and Kr_y are brought close from to 1, as the deviation becomes more apparent (more specifically, as the absolute values of the Vover_x and Vover_y shown in FIG. 18 become larger).

[0318] In this case, each i-th gain coefficient Ki_x (i=1, 2, 3) calculated by the equation 09x approaches from the constant number Ki_a_x on the minimum side to the constant number Ki_b_x on the maximum side, as Kr_x approaches to 1. This applies to each i-th gain coefficient Ki_y (i=1, 2, 3) calculated by the equation 09y.

[0319] Thereafter, by increasing the absolute value of the above-mentioned gain coefficient, the sensitivity towards the manipulated amount (the imaginary wheel rotational angular acceleration commands ωwdot_x_cmd and ωwdot_y_cmd) with respect to the change in the tilt of the base body 9 and the seat 3 (the change in the base body tilt angle error measured value θbe_xy_s), and the change in the estimated center-of-gravity velocity values Vb_xy_s increase. Therefore, when the tilt amount of the base body 9 and the seat 3 from the basic posture tends to increase, or the magnitude of the estimated center-of-gravity velocity values Vb_xy_s tend to increase, the moving velocity of the wheel assembly 5 is controlled so as to promptly dissolve the same. Therefore, it becomes possible to strongly restrict the base body 9 from tilting greatly from the basic position or the estimated center-of-gravity velocity values Vb_xy_s from increasing, and consequently, it becomes possible to restrict one or both of the moving velocity of the wheel assembly 5 in the X-axis direction and the Y-axis direction becomes excessive velocity so as to make one or both of the rotational angular velocities of the electric motors 31R and 31L from deviating from the allowable range at the limiting processor 86.

[0320] (Effect of the Invention)

[0321] According to the omnidirectional vehicle 1 of the present invention, by coordinate converting at least one of “the past estimated value” and “the instantaneous provisional value” of the two-dimensional floor inclination using the orientation of the vehicle 1 (Z component of the base body tilt angle in the world coordinate system θb_z_s_world) or a rate of change thereof (θb_z_s_world−θb_z_s_p_world or θb_z_s_p_world−θb_z_s_world), the coordinate system of the past estimated value and the instantaneous provisional value of the two-dimensional floor inclination are conformed (refer to FIG. 11: the vehicle-world coordinate converter 242 and the world-vehicle coordinate converter 246; FIG. 12: the previous-current vehicle coordinate converter 255; FIG. 13: the vehicle-world coordinate converter 262 and the world-vehicle coordinate converter 268, FIG. 14: the previous-current vehicle coordinate converter 315 and the previous-current vehicle coordinate converter 325; and FIG. 15: the current-previous vehicle coordinate converter 335 and the previous-current vehicle coordinate converter 345).

[0322] Thereafter, by calculating the weighed mean value or the low-pass filter value using the past estimated value and the instantaneous provisional value of the two-dimensional floor inclination, a new estimated value of the two-dimensional floor inclination is calculated (refer to FIG. 11: the calculator 244; FIG. 12: the calculators 251-253; FIG. 13: the calculators 264 and 266; FIG. 14: the calculators 311-313 and 321-323; and FIG. 15: the calculators 331-333 and 341-343).

[0323] As a result, even when at least either one of the driving force of the actuator 7 or the propelling force of the traveling motion unit 5 and the motional states of the vehicle 1 change to an extent that the instantaneous provisional value of the two-dimensional floor inclination changes abruptly, it becomes possible to avoid the situation where the changing amount of the new estimated value with respect to the past estimated value of the two-dimensional floor inclination becomes excessive. Therefore, the operation of the actuator 7 may be controlled appropriately from the viewpoint of stabilizing the posture of the vehicle 1 according to the inclination state of the floor surface, on the basis of the new estimated value of the two-dimensional floor inclination.

[0324] Further, the operation of the actuator 7 is controlled so that the moving form of the vehicle 1 is adjusted appropriately from the viewpoint of stabilizing the posture of the vehicle 1 in the first direction (X-axis direction) and the second direction (Y-axis direction), respectively. To be more specific, in the case where the floor surface is a slant surface in the first direction and the second direction, respectively, the tilting posture of the base body 9 taking the vertical line as a reference is corrected to the uphill side of the slope (the higher surface side) (refer to FIG. 16(c)). By doing so, the tendency of the base body 9 tilting to the downhill side of the slant surface by its own weight is alleviated or relieved, so that stabilization of the posture of the base body 9 in the first direction and the second direction, respectively, and consequently the stabilization of the posture of the vehicle 1 is achieved.

[0325] It may be set so that the operation of the actuator 7 is controlled so as to displace the center-of-gravity 60_—xy of the vehicle 1 (vehicle system center-of-gravity point) in an amount corresponding to the tilt angle θb_xy_s of the base body 9 in the direction to which the base body 9 is tilting, and correct the tilt angle θb_xy_s of the base body 9 so as to generate movement of center-of-gravity of the vehicle 1 equivalent to the displacement amount of the ground contact surface of the traveling motion unit 5 generated by the inclination of the floor surface.

[0326] According to the vehicle 1 of the above-described configuration, in the case where the floor surface is a sloped surface, the tilt posture θb_xy_s of the base body 9 is corrected to the uphill side of the slant surface as is mentioned above, and as a result the motion of the actuator 7 is controlled so as to make the center-of-gravity of the vehicle 1 move towards the uphill side of the slant surface. By doing so, the displacement amount of the ground contact surface towards the downhill side of the slope surface by the weight of the vehicle 1 is cancelled, so that the stability of the position of the vehicle 1 in the direction along the slope surface is achieved, and consequently the stability of the posture of the vehicle 1 is achieved.

[0327] (Other Embodiments of the Present Invention)

[0328] Next, some variations with respect to the embodiments explained above will be explained.

[0329] In the above-explained embodiments, the vehicle system center-of-gravity point (more specifically, the vehicle-occupant overall center-of-gravity point) is set as the predetermined representative point of the vehicle 1. However, the representative point may be set to points such as, for example, the center point of the wheel assembly 5, or a point in a predetermined region of the base body 9 (for example, the support frame 13).

[0330] Moreover, in the above-explained embodiments, the gain coefficients K1_—x, K3_—x, K1_—y, and K3_—y are constantly determined to be |K3_—x|/|K1_—x|<|K3_—y|/|K1_—y|. However, in the present embodiment, in the normal case where one or both of the rotational angular velocities of the electric motors 31R, 31L does not become too fast (in detail, in the case where the output values Vw_x_lim1 and Vw_y_lim1 are not forcibly restricted by the limiting processor 86 of the gain adjustor 78), the gain coefficients K1_—x, K3_—x, K1_—y, and K3_—y respectively coincides with the constant values K1a_x, K3_—a_x, K1a_y, and K3_—a_y. As such, for example, it may be set so that they are determined to be |K3_—x|/|K1_—x|<|K3_—y|/|K1_—y|, and consequently |Vb_x_stb|/|θbe_x_s|>|Vb_y_stb|/|θbe_y_s| only in the normal case. In this case, the constant values K1a_x, K1a_y, K3_—a_x, and K3_—a_y should be set so as to be |K3_—a_x|/|K1a_x|<|K3_—a_y|/|K1a_y|.

[0331] In contrary to above, the gain coefficients K1_—x, K3_—x, K1_—y, and K3_—y may arbitrarily be determined so as to be |K3_—x|/|K1_—x|>|K3_—y|/|K1_—y|, according to the operating state or operating environment of the vehicle 1, or to a switching operation by the occupant or the like. In this case, it becomes possible to restrain the traveling direction of the wheel assembly 5 from dispersing with respect to the Y-axis direction, in the case where the vehicle 1 is made to travel in the Y-axis direction.

[0332] Further, in the above-explained embodiments, the gain coefficient Ki_xy (i=1, 2, 3) is made to arbitrarily change by the gain adjustment parameter Kr_xy. However, it may be set so that the gain coefficient Ki_xy (i=1, 2, 3) is maintained stable in both or either one of the boarding mode and the autonomous mode. In the case where the gain coefficient Ki_xy (i=1, 2, 3) is maintained stable, the processing of the gain adjustor 78 will not be necessary.

[0333] Further, in each of the aforesaid embodiments, the vehicle 1 having the structure illustrated in FIG. 1 and FIG. 2 has been exemplified. However, the omnidirectional vehicle 1 in the present invention is not limited to the vehicle 1 exemplified in the above-explained embodiments.

[0334] To be more specific, the wheel assembly 5 serving as the traveling motion unit of the vehicle 1 in the above-explained embodiments has the one-piece construction. Alternatively, however, the wheel assembly 5 may have a construction shown in, for example, FIG. 10 of the aforesaid patent document 3. More specifically, the wheel assembly may be constructed to have a rigid annular shaft member and a plurality of rollers rotatively and externally inserted into the rigid annular shaft member such that the axial centers thereof are oriented in the tangential direction of the shaft member, the plurality of these rollers being arranged in the circumferential direction along the shaft member.

[0335] Further, the traveling motion unit may have a crawler-shaped structure, as shown in, for example, FIG. 3 of patent document 2.

[0336] Alternatively, as shown in, for example, FIG. 5 of the aforesaid patent document 2, FIG. 7 of patent document 3, or FIG. 1 of patent document 1, the traveling motion unit may be constructed of a spherical member, and the vehicle may be constructed such that the spherical member is rotatively driven in a direction about the X-axis and a direction about the Y-axis by an actuator (e.g., an actuator having the aforesaid wheel assembly 5).

[0337] Further, in the above-explained embodiments, the vehicle 1 provided with the seat 3 as the payload supporting part for an occupant has been exemplified. Alternatively, however, the omnidirectional vehicle in accordance with the present invention may be a vehicle having a construction wherein a step on which an occupant rests his/her both feet and a section to be gripped by the occupant standing on the step are mounted on a base body, as illustrated in, for example, FIG. 8 in patent document 3.

[0338] Thus, the present invention can be applied to omnidirectional vehicles of various constructions, as illustrated in the aforesaid patent documents 1 to 3 and the like.

[0339] Further, the omnidirectional vehicle in accordance with the present invention may be provided with a plurality of traveling motion units (e.g., two in the lateral direction, or two in the fore-and-aft direction, or three or more) capable of traveling in all directions on a floor surface.

[0340] Further, in the omnidirectional vehicle in accordance with the present invention, it is not necessary for the base body to tilt together with the payload supporting part of the occupant. For example, in the case where the vehicle includes a plurality of the traveling motion units, the base body, which the traveling motion units are assembled to, may be made so that the base body does not tilt with respect to the floor surface, and the payload supporting part may be assembled to the base body so as to be freely tiltable.