Biodynamic Resistant Control Stick

1. Field of the Invention

This invention generally relates to aircraft control systems, and in particular to pilot operable members of such control systems.

2. Description of the Prior Art

It is well recognized that aircraft control systems have become increasingly complex. This complexity is due to the requirement that aircraft, particularly tactical military aircraft, perform over a wide speed regime. Control system complexity has also increased due to a requirement for redundancy with so called "fly by wire systems" wherein there is no direct mechanical linkage between the pilot operated control member and the aircraft control surfaces.

Such control systems have included means for varying the force required to move the pilot operated member in response to the forces acting on the aircraft or the speed of the aircraft. Such force varying components and dampers have been connected in series between the pilot operated control member and the aircraft control surfaces so as to increase aircraft stability.

Additionally, high speed tactical aircraft tend to become unstable during certain speed regimes. In such cases, electronic stability augmentation systems are added to the aircraft control system so as to provide the requisite degree of inherent stability. Pilot induced oscillations, which tend to occur during certain speed regimes, are routinely compensated for by such stability augmentation systems. The electronic stability augmentation systems of the prior art are placed in series between the pilot operable member or control stick and the aircraft control surfaces.

It has long been recognized that movement of an aircraft may inadvertantly cause the pilot to move the control stick, thereby causing the aircraft to deviate, at least momentarily, from its desired flight path. A well known example of this "biodynamic feedthrough" occurs when an aircraft encounters turbulent air. The aircraft may be caused to pitch suddenly thereby causing the pilot to either pull or push the control stick. This biodynamic action causes further pitch excursions of the aircraft. Although this biodynamic feedthrough has been a problem to varying degrees, it has generally been handled by recognition of the problem by the aircraft pilot and the resulting appropriate pilot training. However, with advancing aircraft technology, such inadvertant deviations from the aircraft's desired flight path have become unacceptable.

For example, certain high technology, military tactical aircraft are now equipped with fuselage mounted deflectors which, when extended on one side of the aircraft, cause lateral movement of the aircraft in the opposite direction. Such lateral accelerations inadvertantly cause the pilot to laterally move the aircraft's control stick, thereby causing the aircraft to undesirably roll or rotate about its longitudinal axis.

It is therefore an object of the present invention to provide a means for preventing intentional movement of an aircraft in one mode to unintentionally cause movement of the aircraft in another mode.

It is another object of the present invention to provide a means for causing an aircraft pilot operated member to resist movement thereof caused by the pilot due to movement of the aircraft.

According to the present invention, a means is provided for resisting biodynamic feedthrough. A spring is connected at its midpoint to the aircraft control stick. An actuator is connected to the opposite ends of the spring so as to compress it. The spring resists movement of the stick from a neutral position, at all times. A force sensor is provided which senses lateral movement of the aircraft and generates a signal which causes the actuator to compress the spring. Thus, the force required to move the stick, from its neutral position, is increased so as to reduce the amount of biodynamic feedthrough induced by the lateral movement of the aircraft.

In another embodiment of the invention, a force is applied to the control stick so as to counter the biodynamic feedthrough. This may be accomplished by a spring having one end connected to the actuator and its opposite end connected to the control stick, so that movement of the actuator exerts a force counter to the biodynamically induced force. The same result may be achieved by a spring connected to the control stick at its midpoint, having one end connected to the actuator and its other end connected to the aircraft frame.

Under certain circumstances, a damper may be desired. In such circumstances, the force sensor would provide a signal to a damping device connected to the stick and to the aircraft frame, so as to vary the degree of damping in accordance with lateral acceleration of the aircraft.

The objective of both embodiments of the invention is to resist the biodynamically induced force exerted by the pilot on the control stick. The force or acceleration sensor provides an input signal which varies directly with the lateral acceleration of the aircraft. This signal is processed by a processing means which generates a processed signal which causes the actuator to provide the appropriate degree of resistance. The correlation between lateral acceleration and the degree of resistance required at the control stick may be determined experimentally or from a variety of algorithms which vary depending upon the particular application of the device. The processor generates a process signal in accordance with that correlating information, and may take the form of an analog circuit or a microprocessor.

The construction of the preferred embodiment, as well as further objects and advantages of the invention, will become apparent from the following specification when considered with the accompanying drawings in which like numerals refer to like parts wherein:

FIG. 1 shows, schematically, an aircraft.

FIG. 2 schematically shows an aircraft control stick incorporating the present invention.

FIG. 3 schematically shows an alternative embodiment of a portion of the aircraft control stick assembly of the present invention.

FIG. 4 schematically shows an analogy of the human body and control stick.

FIG. 5 schematically shows a side view of an aircraft control stick and human arm.

FIG. 6 is a front view of the control stick and arm shown in FIG. 5.

FIG. 7 shows a pilot for the aircraft shown in FIG. 1.

FIG. 8 shows a plot of H versus K.sub.s /M.sub.s versus B.sub.s /M.sub.s.

FIG. 9 is a flow diagram of an algorithm which may be used with the invention.

FIG. 1 shows an airplane 2 and its three orthogonal axes. The generally recognized sign convention is used. The airplane 2 is shown with an outwardly deflectable canard 4. The canard 4 is normally flush with the aircraft's fuselage. The canard is pivotally mounted at its leading edge so that it may be deflected outward. When the canard 4 is deflected, a sideward force F.sub.1 is aerodynamically generated which causes the aircraft to accelerate and move laterally. FIG. 1 shows such a canard on the right side of the aircraft, such that when it is deflected, the aircraft is caused to move laterally to the left.

The canard 4 is extended by an actuator, not shown. Such canards are typically mounted on both sides of the aircraft and are positioned so that the lateral force F.sub.1 acts at the aircraft's center of gravity. Two canards may be used on each side of the aircraft, in which case they are positioned so that their combined effect is to produce the force F.sub.1 so that it acts at the aircraft's center of gravity. Such devices enable military aircraft to quickly engage in evasive manuevers. The canards on each side of the aircraft may be operated simultaneously so as to act as speed brakes. In such a case, the two laterally directed forces which are generated by the respective pair of canards cancel one another and no lateral movement of the aircraft occurs.

Sudden accelerations of an aircraft result in an opposite, but equal reaction, on persons and objects within the aircraft. For example, in the example shown in FIG. 1 the aircraft is thrust in the positive y direction, or to the left, when the canard 4 on the right side of the aircraft is deflected. This causes the pilot to be forced to the right side of the aircraft.

Reference will now be made to FIG. 2 where a rear view of a schematic of an aircraft control stick assembly is shown. A control stick 6 is shown being gripped by a pilot's right hand 8. The control stick assembly which is schematically shown is of the type known as a side arm controller, and is shown applied to a fly by wire system. Typically, such a side arm controller is a stick which is mounted on the right side of the pilot. Movement of the stick 6 causes a signal to be sent to electrical actuators which operate the aircraft's control surfaces. Such actuators may be electrical or hydraulic, and the signal which is generated by movement of the control stick may be electrical or hydraulic. For purposes of explanation, the control stick 6 is pivotally mounted to the aircraft frame 3 by a ball and socket assembly 10 which permits lateral and longitudinal motion of the stick 6 about its pivot point. Typically, lateral movement of such a control stick causes the aircraft to rotate about its longitudinal or x axis through the deflection of ailerons or spoilers. Longitudinal movement of the stick typically causes rotation of the aircraft about its pitch or y axis through deflection of its elevators.

As may be seen in FIG. 2, lateral movement of the stick 6, along the y axis, causes the stick to be pivoted about its ball and socket so that a rack 12 causes rotation of a gear 14 which is connected to a potentiometer 15, thereby generating a signal which is used to operate the actuators which control the aircraft's ailerons or spoilers so as to cause the aircraft to rotate about its roll or x axis.

When the canard 4 is caused to deflect by the pilot, by means not shown, the force F.sub.1 is generated on the aircraft which causes the aircraft to accelerate laterally in the positive y direction. The pilot's right arm and hand 8 is thereby acted upon by a force F.sub.2 and generates a moment M.sub.2, which tends to move the stick to the right, thereby causing the aircraft to roll to the right, or in the negative R.sub.x direction. This action is commonly referred to as biodynamic feedthrough, and is undesirable in that the pilot's objective in deflecting the right canard was to move the aircraft to the left. The pilot did not intend the aircraft to roll to the right, but was caused to induce the right hand roll due to the force field which was generated by the lateral acceleration of the aircraft to the left.

An apparatus embodying the principles of the present invention and used to counter the effects of F.sub.2 will now be described. A spring 16 is shown connected at its midpoint to the control stick 6 below the ball and socket 10. The opposite ends of the spring 16 are connected to the arms 18 of a turn buckle 19. A screw 20 of the turn buckle 19 is caused to rotate by a motor 22 connected to the aircraft frame 3. Thus, rotation of the motor will cause the spring to be compressed or stretched, thus varying the spring constant so as to vary resistance of the control stick 6 to movement from a neutral position. The spring 16 may be nonlinear (have a different number of turns per inch) to achieve this effect. The motor 22 rotates in response to a signal received from a processor 24. The processor 24 transmits the signal to the motor based upon a signal it receives from an accelerometer 26.

In operation, deflection of the canard 4 produces a force F.sub.1 which moves the aircraft to the left, in the +y direction. This, in turn, causes the pilot to unintentionally exert a force F.sub.2 on the control stick 6. Additionally, F.sub.1 induces a force F.sub.3 to act on the accelerometer 26 which transmits a signal to the processor 24 which, in turn, sends a signal to the motor 22 which causes the screw 20 to rotate, thereby drawing arms 18 toward each other. The drawing together of the arms compress the spring 16, increasing its spring constant K.sub.s, making it more difficult to move the stick from its neutral position. Movement of the stick is resisted by a force F.sub.4, not shown, which generates a moment M.sub.4 which resists the biodynamic moment M.sub.2. The force F.sub.4 acts at the point at which the spring 16 is connected to the stick. The actual magnitude of F.sub.4 should be such that the restoring moment M.sub.4 is substantially equal to the biodynamic moment M.sub.2. In the present system, F.sub.2, and, hence, the biodynamic moment M.sub.2 tending to rotate the stick, is a direct function of F.sub.1. The force F.sub.3, which acts on the accelerometer, is also a direct function of F.sub.1. The objective is to use the signal generated by the accelerometer 26 to appropriately compress the spring 16 so that the force F.sub.4, and, hence, the restoring moment M.sub.4, is substantially equal to the biodynamically induced moment M.sub.2. It is the function of the processor 24 to produce a process signal in response to the signal received from the accelerometer 26, which will produce the appropriate degree of compression of the spring 16 so as to generate the proper restoring moment M.sub.4.

In general, for small deflections of the control stick, in the example described, there is a linear relationship between M.sub.4 and M.sub.2. Accordingly, with appropriate calibrations, the processor 24 could take the form of a simple analog electrical circuit. However, for large deflections of the control stick, the relationship between M.sub.2 and M.sub.4 becomes non-linear. If the degree of non-linearity is unacceptable, a more complicated circuit could be substituted. Such electrical circuit may include a voltage divider or wheatstone bridge.

A statistical or empirical data base may also be obtained, by testing, to determine the performance characteristics of the processor 24. Once such a experimental data base has been obtained, it may be incorporated either in an appropriate analog electrical circuit, or in a microprocessor which would extrapolate between the experimentally determined data points.

A preferred approach for determining the performance characteristics for the processor 24 is to calculate the desired characteristics and program an equation, or an algorithm based upon the equation in a microprocessor. An example of such an equation and algorithm is set forth in an appendix to this patent application.

The particular processor, i.e., electrical analog circuit, microprocessor, or other device, as well as the performance characteristics of the processor 24, will vary depending upon the particular application of the invention and various other design criteria. The processor 24 may also have adjustment means to take into account various other factors. Such other factors would include pilots having arms of varying masses, which would result in a change in F.sub.2 for a particularly F.sub.1. Another factor which would lend itself to an adjustment would be aircraft control sticks which have adjustable lengths. In the side arm controller, used for the application described in the present case and in the algorithm set forth in the attachment, it was assumed that the pilot's elbow was fixed and that the pivoting motions of the hand and control stick would take place about a pivot point located at the elbow. The relationship of F.sub.2 to F.sub.1, and, hence, the relationship between M.sub.4 and F.sub.1 change considerably if the pilot's elbow is not fixed. This change would also have to be taken into account in the performance characteristics of the processor 24.

Another embodiment of the invention is shown in FIG. 3. In this embodiment, the spring 16 is connected at its midpoint to the control stick 6. One end of the spring 16 is connected to the aircraft frame 3, while the other end is connected to a rod 40 which is caused to move parallel to the spring center line by a gear 42 which is driven by a motor or actuator 44. The motor 44 rotates in response to a voltage signal proportional to the desired change in the spring constant received from processor 24. In this embodiment of the invention, only one end of the spring is moved by the motor 44. Thus, compression of the spring, in addition to changing the spring constant, exerts a restoring force and, hence, restoring moment on the control stick 6, without deflection of the stick by the biodynamically induced force F.sub.2. By eliminating that portion of the spring between the control stick and the aircraft, a restoring force F.sub.4 and restoring moment M.sub.4 may also be applied to the control stick. This method of opposing the biodynamically induced force F.sub.2 may be desirable under certain circumstances and for certain applications.

In the application of the invention described herein, it was found that little or no damping of the stick was required due to the biodynamic input F.sub.2. However, there are certain circumstances under which damping may be desirable. In such an event a damping device may be provided as shown in FIG. 2. A link and piston 28 is slidably connected to a cylinder 30. A second link and piston 32 is also slidably connected to the cylinder 30. The cylinder contains a compressible fluid. The link 32 is connected to a motor 34 which receives a signal from the processor 24. Accordingly, movement of the link 32 varies the pressure of the fluid within the cylinder 30, thereby varying the damping characteristics or the damping constant B.sub.s between the control stick 6 and the aircraft frame 3.

The present invention has been applied and described with reference to a coupling of aircraft roll to aircraft side force. It will be apparent to those skilled in the art that the present invention may be applied to any one of a number of similar aircraft parameters. For example, turbulence frequently cause aircraft to pitch about its pitch or y axis, thereby causing the pilot to move longitudinally in a direction parallel to the aircraft's roll axis. Such motions result in a biodynamic input to the control stick, which cause the aircraft to further pitch about its roll or y axis. The present invention may be adapted so as to sense aircraft motion along the x axis and apply a restoring moment to the aircraft control stick so as to eliminate or substantially reduce biodynamic input which would cause fore and aft motion of the control stick. The invention has been described with respect to a fly by wire side arm controller. The invention could be equally applied to a conventional mechanical aircraft control stick or yoke which is centrally located with respect to the pilot.

Although the present invention has been described with reference to a particular embodiment herein set forth, it is understood that the present disclosure has been made only by way of example and that numerous changes in the details of construction may be resorted to without departing from the spirit and scope of the invention. Thus, the scope of the invention should not be limited by the foregoing specification, but rather only by the scope of the claims appended hereto.

The approach to solving the problem can be schematically represented as shown in FIG. 4 where M.sub.1 is the human body and the augmented system shown in the box is the control stick of the present invention. By approximately selecting the spring constant K.sub.2, movement of M.sub.1 need not result in movement of M.sub.2. The sign convention used herein is shown in FIG. 7 with respect to an aircraft pilot who will use a side arm controller. FIG. 5 is a schematic illustration of the pilot's right arm, from the elbow forward, and assumes that the elbow is stationary at point A. The pilot is grasping the side arm controller or control stick. FIG. 6 is a front view of the arm-stick combination shown in FIG. 5.

For illustrative purposes, this appendix will describe the derivation of an algorithm for minimum biomechanical feedthrough. After the algorithm is derived, the microprocessor in the stick controller can change the stick properties (gain K*(t) and damping ratio B*(t)) so that the total man-machine interaction has minimum biomechanical feedthrough. This choice of the objective function (biomechanical feedthrough) being minimized is somewhat arbitrary. Any other type of function could, alternatively, be minimized. This would produce a different type of dynamic stick which may have application in other situations.

For the minimum feedthrough stick, it can be shown that the configuration shown in FIGS. 5 and 6 produce equations of motion that decouple under small angle assumptions. Since the G.sub.y forces act in only the y-z axis in FIG. 7, the equations of motion are derived here only in this axis. The Lagrangian for motion in the x-y plane can be expressed as: ##EQU1## where an assumption of small angular rotations of stick and arm, respective mass concentration at the midlength of the stick and arm, and an elastic coupling of magnitude K.sub.I between the hand and stick have been made. The equations of motion are then directively derivable from: ##EQU2## where q.sub.i =x.sub.3 or x.sub.4, and where, for the forearm:

and the stick:

and F.sub.1 (t) and F.sub.2 (t) are respectively applied external forces to the forearm and stick, and B.sub.s is the damping coeficient for stick motion. The resulting equations are:

These equations can be written in state variable form as follows: ##EQU3## where the choice of state variables is:

Typical values of the relevant constants in FIGS. 5 and 6 are displayed in Table I. These measurements were made in the vibration theory area and are summarized in (2). The optimization problem of interest can now be formulated such that it is possible to design K.sub.s, B.sub.s, and M.sub.s to achieve an optimal design.

Formulation of the Optimization Problem:

From FIGS. 5 and 6, the problem of interest is to pick values of K.sub.s, B.sub.s, and M.sub.s in such a manner as to minimize deflections at x.sub.3 (t) due to disturbances from F.sub.2 (t) which are a result of a G.sub.y force vector field. It is noted that when the optimal values of K.sub.s, B.sub.s, and M.sub.s (denoted as K.sub.s *, B.sub.s *, and M.sub.s *) are obtained, these values may differ from the true stick values. The differences:

become the values of the augmented system which must be added to the true stick parameters to improve the total response of the man-machine interaction. This is analogous to the vibration resistant problem in which the augmented system is designed to produce zero amplitude of the mass M.sub.1 in FIG. 4. The difficulty problem at this point is in the determination of the criteria of optimality. If is is assumed that F.sub.2 (t) can be decomposed into its Fourier components: ##EQU4## where n is fixed. For small angles, the system is decoupled between the z-x plane and the z-y plane, then equation (7) can be written:

where

is the only term which depends on the parameters to be optimized. If the definition of optimality is such as to produce minimal amplitude at x.sub.3 (t) due to F.sub.2 (t), then J may be minimally defined as follows: ##EQU5## where .DELTA.T is the duration of the tracking task, T indicates matrix transpose, and Q is a positive semidefinite matrix. If: ##EQU6## then the optimization problem consists of the minimization of only the measurement variables x.sub.3 (t) and x.sub.4 (t). If, however, Q is given by: ##EQU7## then the cost functional J will include minimization of the variables x.sub.3 (t), x.sub.3 (t), x.sub.4 (t) and x.sub.4 (t). The optimization problem can be summarized as follows:

Minimize J(K.sub.s, M.sub.s, B.sub.s) with respect to the variables K.sub.s /M.sub.s, and B.sub.s /M.sub.s subject to equation (12), (10), and F.sub.2 (t) specified in equation (9) (assume M.sub.s =contant for simplicity).

Solution of the Optimization Problem

The optimization problem specified by J of equation (12) with Y(t) satisfying (10) and F.sub.2 (t) satisfying (9) is a problem of the Lagrange type [3]. To solve this problem, the Hamilitonian of the system is defined by:

where .lambda..sup.T (t)=[.lambda..sub.1 (t), .lambda..sub.2 (t)] is the Lagrange multiplier vector. The Euler equations of optimality for an optimal solution are of the form:

and the third equation to be satisfied for optimality is: ##EQU10## where K=(K.sub.s /M.sub.s, B.sub.s /M.sub.s)=the parameters for which the optimization is conducted. If (20) is a linear function of the parameters, (20) is replaced by the condition that ((K.sub.s /M.sub.s)*, (B.sub.s /M.sub.s)*) be determined such that:

for all admissible values of (K.sub.s /M.sub.s) and (B.sub.s /M.sub.s). This occurs in the case that some range of values of (K.sub.s /M.sub.s) and (B.sub.s /M.sub.s) are known, e.g.

is the range of allowable values.

Since:

or

Then partial derivates of H in equation (24) with respect to K.sub.s /M.sub.s and B.sub.s /M.sub.s would not yield a linear solution for the parameters. This is due to the fact that H is not quadratic in the parameters. Using the relationship (21), the minimization of H must be conducted over a grid of values. This is analogous to final time problems [3] in which the state variables of the .lambda. equation are linear in the variables to be optimized. The three necessary conditions for optimality can now be summarized as follows:

for all admissible values of (K.sub.s /M.sub.s) and (B.sub.s /M.sub.s) which fall within a constraint set specified by equations (22a-b). An algorithm to determine the optimal gains can be stated as follows:

Step 1: Pick an initial value of (K.sub.s /M.sub.s).sub.o and (B.sub.s /M.sub.s).sub.o.

Step 2: Integrate Y(t) foward using (25a-b).

Step 3: Integrate .lambda.(t) backwards using (26a-b).

Step 4: Evaluate H.sub.i (K.sub.s /M.sub.s).sub.i) for this iteration.

Step 5: Pick a different value of (K.sub.s /M.sub.s).sub.i+1 and (B.sub.s /M.sub.s).sub.i+1 from the grid of values and go to step 2.

Step 6: Repeat steps 2-5 for a grid of values unit the optimal values (K.sub.s /M.sub.s)* and (B.sub.s /M.sub.s)* are obtained such that the relationship (27) is satisfied.

It is noted that a gradient procedure could be used to determine the parameters which minimize H in lieu of a grid approach. For purposes of understanding the problem better, the grid method is chosen based on representative values displayed in Table I.

Using a grid method, the variable H (which is a measure of the feedthrough) is plotted for different values of K* and B* (M*=0.4 Kg.=constant) for different constant values in FIG. 8. It is obvious from this diagram that the minimum amount of stick feedthrough occurs for a fixed (and unique) value of K* and B* specified. FIG. 8 was obtained using the typical values of human measurements in Table I with .DELTA.T specified at 100 seconds. Computer Attachment I illustrates the program which gives rise to the plot in FIG. 8.

For the dynamic simulation K*(t) and B*(t) vary with time and the entire algorithm (steps 1-6) must be used. In this case it is required to integrate Y(t) forward in the microprocessor, integrate .lambda.(t) backwards in the microprocessor and to use some type of iteration procedure to update the unknown initial conditions .lambda.(t.sub.o). Computer Attachment II illustrates the dynamic program which changes the gain K*(t) as a function of time and the dashpot value B*(t) as a function of time. For a dynamic simulation of the same G.sub.y data and constants used in FIG. 8 (the only change was the time varying quantities K*(t) and B*(t)), the value of H determined from the simulation was: H=0.00268.

The updating procedure used in Computer Attachment II to iterate on the initial conditions .lambda.(t.sub.o) was a gradient method based on the principal: ##EQU11## FIG. 9 illustrates a flow chart of the total algorithm used in Computer Attachment II. It should be emphasized that the computer algorithm could be used to change the stick's dynamical characteristics (K*(t), B*(t)) in any manner desired by simply changing the microprocessor algorithm which can be done quite easily.

[1] Jex, Henry, R., and Raymond E. Magdaleno, "Biomechanical Models for Vibration Feedthrough to Hands and Head for Semisupine Pilot", Aviation, Space, and Environmental Medicine, Vol, 49, No. 1, Sec. II, January, 1978.

[2] Jex, H. R. and R. E. Magdaleno, "Modeling Biodynamic Effects of Vibration", Final Scientific Report No. 1037-1, Systems Technology Inc., August, 1979.

[3] Bryson, A. E. and Y. C. Ho, "Applied Optimal Control", Ginn & Company, 1969.

[4] Athans, M. and F. C. Schweppe, "Gradient Matrices and Matrix Calculations", MIT Lincoln Laboratory, Lexington, Mass., Tech. Note 1965-53, November, 1965. ##SPC1## ##SPC2##