Rotation Rate Measuring Instrument

The present invention relates to a rotation rate measuring instrument. An instrument of this kind is described in an article by J. L. Davis and S. Ezekiel, "Techniques for Shot-Noise-Limited Inertial Rotation Measurement Using a Multiturn Fiber Sagnac Interferometer", Proceedings of the Society of Photo-Optical Instrumentation Engineers, Volume 157, Laser Inertial Rotation Sensors, Aug. 30-31, 1978, San Diego, Calif. pages 131-136.

The object of the invention is to provide a rotation rate measuring instrument which permits high-accuracy measurements.

With the solution described herein a highly stable control signal is obtained for maintaining the phase difference between the component beams travelling along the closed light path in opposite directions at a constant value. The rotation rate can also be determined directly from the amplitude of the signal from which the control signal is derived.

A development of the invention makes it possible to allow for variations of the scale factor (the scale factor gives the relationship between the phase difference between the two component beams due to the Sagnac effect and the rotation rate).

In another embodiment, it is possible in a simple manner to expand the measuring instrument so that the rotation rate can be measured about three mutually perpendicular axes.

In the embodiment in which the light paths are formed by optical waveguides embedded in a sealing compound, the influence of disturbing effects on the quantities entering the measurement is small.

The invention will now be explained in more detail and by way of example with reference to the accompanying drawings, in which:

FIG. 1 is a block diagram of the rotation rate measuring instrument; FIGS. 2a-2f show diagrams serving to explain the operation of the instrument of FIG. 1, and

FIG. 3 is a block diagram of a development of the instrument of FIG. 1.

First, the design and operation of the rotation rate measuring instrument will be described with the aid of FIG. 1. A laser 1 generates a coherent light beam of frequency F.sub.L, which passes through a Bragg cell 2, where it is modulated at the frequency f(mod). Through this frequency modulation with the constant value f.sub.(mod), the frequency of the output signal of the Bragg cell 2 is F.sub.o =F.sub.L +f(mod).

The purpose of this Bragg cell is to prevent any light reflected to the laser 1 from interfering with the laser. Through the use of the Bragg cell 2, the reflected light has a frequency different from that of the light emitted by the laser, whereby any interference with the laser is prevented. After traversing the Bragg cell 2, the light beam passes through a first beam splitter 3 and a single-mode filter 4 to a second beam splitter 5. The second beam splitter 5 splits the light beam L into two component beams L(CW) and L(CCW), which are coupled into a closed light path 7 and travel along this path clockwise L(CW) and counterclockwise L(CCW), respectively.

One of the two component beams (in the FIG. 1 block diagram the component beam L(CW), which traverses the closed light path clockwise) is frequency-modulated in an additional Bragg cell 6. The choice of the modulation frequencies will be explained later.

The closed light path can be implemented in various ways, e.g. with suitably arranged mirrors. To obtain a long optical path length, the mirrors are arranged so that the light path has approximately the shape of a coil spring whose individual turns have the same radius.

In the embodiment being described, the closed light path is formed by an optical waveguide 7 arranged in the form of a coil. The radius of a turn is R. If the coil has m turns, the length of the entire light path is L=2.pi.m.

After emerging from the optical waveguide, the component beam L(CW) returns to the second beam splitter 5 direct, while the other component beam L(CCW) passes through the Bragg cell 6 before returning to the second beam splitter 5.

The beam splitters 3, 5 are designed to pass a first portion of the light beam impinging on them, to deflect a second portion through 90.degree., and to reflect a further portion which is small compared with the first and second portions.

The second beam splitter 5 is struck by two beams, namely the component beams L(CW) and L(CCW), after these beams have traversed the optical waveguide 7. These two component beams are superimposed on each other in the beam splitter 5, from which issue two mutually perpendicular light beams each resulting from the superposition of the two component beams. Their intensities are I.sub.1=k.multidot.cos.sup.2 .phi.s/2 and I.sub.2 =k.multidot.sin.sup.2 .phi.s/2, where k is a constant, and .phi..sub.s is the phase shift between the two component beams due to the Sagnac effect, which will be dealt with later.

The light beam of intensity I.sub.1 passes through the single-mode filter 4 to the first beam splitter 3, which deflects it to an evaluating unit 80, whose design and operation will be explained in detail later. The path of this deflected light beam is perpendicular to the path of the light beam travelling from the first Bragg cell 2 to the first beam splitter 3.

For beam guidance and for coupling the beams into and out of the individual components and the optical waveguide, further optical beam-focussing and imaging devices are required. These are familiar to those skilled in the art and have been omitted in the block diagram to simplify the illustration.

The beam splitters used may be of the same kind as those employed in the interferometer disclosed in the above reference, and a suitable laser is a HeNe laser. The single-mode filter is a single-mode optical waveguide 10 m in length. The optical waveguide 7 is a polarization-conserving single-mode optical waveguide.

If no polarization-conserving optical waveguide is used, further optical devices are necessary for polarization control and measurement. Suitable Bragg cells are those described in a publication by Soro, 26 rue Berthollet, 94110 Arcueuil, France, entitled "Modulateur OEM Acousto-Optique", Bulletin MD811. All devices employed are known per se.

The optical devices can be implemented wholly or in part using integratedoptic technology.

With Bragg cells (i.e., acoustooptical modulators), frequency modulation is possible. One portion of the input beam emerges from the Bragg cell unmodulated at a first angle, while the other portion emerges frequency-modulated (a frequency shift is produced at a constant modulation frequency) at a second angle. The angle of emergence of the frequency-modulated beam depends on the modulation frequency. An improvement in the separation of the modulated and unmodulated beams can be achieved in an advantageous manner by grinding the outside surface of the crystal of the Bragg cell in such a way that the modulated beam is normal to the outside surface.

In the following, the operation of the rotation rate measuring instrument and particularly of the evaluating unit 80 will be described.

If the two component beams propagate along the coil-shaped waveguide of length L in opposite directions, and the coil, whose radius is R, rotates about its axis at the angular velocity .OMEGA., the two component beams emerging from the optical waveguide differ in phase by ##EQU1## where .lambda.=vacuum wavelength of light, and C=velocity of light in vacuo.

The expression 4.pi.L.multidot.R/.lambda.C contains quantities which are constant only in principle, and is called "scale factor". Temperature variations and pressure effects may slightly change L and R, which results in errors in the measurement of .OMEGA..

The evaluating unit 80 contains an optical/electrical transducer 8, whose electric output signal is directly proportional to I.sub.1 =k cos.sup.2 .phi.S/2. FIG. 2a shows how the intensity I.sub.1 varies qualitatively with the phase difference between the two component beams after traversal of the optical waveguide. It can be seen that for small changes of .phi.(and, hence, .phi..sub.S) in the range .phi.=O, the intensity I.sub.1 hardly changes, i.e., the measurement accuracy is low in this range.

An improvement is advantageously obtained by introducing an additional constant phase shift .phi..sub.B1 between the two component beams. .phi..sub.S is then superimposed on .phi..sub.B1. If .phi..sub.B1 =.pi./2, small changes of .phi..sub.S result in a large change of I.sub.1. The phase difference between the two component beams is then .phi.=.phi..sub.S +.phi..sub.B1.

The additional phase shift .phi..sub.B1 is produced by the second Bragg cell 6. To obtain the desired phase shift .phi..sub.B1 between the two component beams, care must be taken to ensure that the Bragg cell is not located in the middle of the light path, i.e., in the middle of the optical waveguide 7, for the effects on the two component beams would then be the same. In the embodiment, the Bragg cell is located at one end of the optical waveguide.

The Bragg cell 6 modulates the component beam L(CW) at the frequency f1, so that the beam propagates through the optical waveguide at the frequency F.sub.o +f1. The other component beam L(CCW) propagates through the optical waveguide at the frequency F.sub.o. At rest, the phase difference .phi..sub.B1 between the two component beams at the ends of the optical waveguide is ##EQU2## where n is the refractive index of the optical waveguide.

In the novel rotation rate measuring instrument, switching is effected between two modulation frequencies f1 and f2 for the Bragg cell 6 and the frequency F.sub.S (FIG. 2b). The phase difference produced by the modulation at f2 is ##EQU3## The difference between the two modulation frequencies f1 and f2 is chosen so that ##EQU4## In this case (if .phi..sub.S =0, i.e., for zero rotation rate), I.sub.1 for .phi..sub.B1 is equal to I.sub.1 for .phi..sub.B2. Thus, no variation occurs at the optical/electrical transducer due to the switching (FIG. 2c).

The above condition for .phi..sub.B2 -.phi..sub.B1 is a special case for the choice of the phase difference between .phi..sub.B2 and .phi..sub.B1. In general terms, the condition for the choice of the modulation frequencies (and, hence, of the phase difference) is as follows: The two modulation frequencies f1 and f2 must be chosen so that I1 has the same value for the phase shifts .phi..sub.B1 and .phi..sub.B2.

Things are different if rotation is taking place, i.e., if the phase difference .phi..sub.S is superimposed on the phase shifts .phi..sub.B1 and .phi..sub.B2. Now, as can be seen in FIG. 2a, I1 has different values depending on the respective modulation frequency (i.e, on the different phase shifts .phi..sub.B1, .phi..sub.B2). This is illustrated in FIG. 2d for clockwise rotation and in FIG. 2e for counterclockwise rotation. As long as the measurement is made in the linear portion of the characteristic, the difference between the maximum and the minimum values of this square-wave intensity curve is proportional to the rotation rate .OMEGA.. This difference in intensity can be evaluated in a simple manner as follows (not shown in the drawings): The output signal of the optical/electrical transducer, which is a square wave, is applied to a band-pass filter whose midfrequency is equal to the switching frequency F.sub.S. The band-pass filter provides a sine-wave signal having an amplitude A proportional to the rotation rate .OMEGA.. .OMEGA. is calculated by the formula

The amplitude measurement is performed in the known manner. The direction of rotation is obtained by comparing the phase of the switching signal F.sub.S with that of the output signal of the band-pass filter.

Instead of directly evaluating the amplitude, it is also possible to control the phase difference between the two component beams L(CW) and L(CCW) so that it is maintained at a constant value. In this case, the control signal is directly proportional to the rotation rate.

The phase difference is preferably held at the values (2p-1).pi./2, because in this case the measurement accuracy is highest for the reasons mentioned above. Then, .phi..sub.B1 +.phi..sub.S =.pi./2 (2p-1). To produce the phase difference (2q-1).pi. between the two component beams after traversal of the optical waveguide, one of the two component beams is frequency-modulated (whereby the phase difference is produced) successively at f1 and f2. f1 is equal to F*.+-..delta.F, and f2 is equal to F*+F.+-..delta.F. F* is the frequency of a VCXO 124 when the control signal for the VCXO is zero. .delta.F is proportional to the rotation rate .OMEGA.. F is chosen so that the desired phase difference of .pi. is produced between the component beams (L(CW) and L(CCW).

This will be illustrated by a numerical example: L=1,400 m; .phi..sub.B1 =6,801 .pi./2;.phi. .sub.B2 =6,803 .pi./2. From this it follows that F*=249.547 MHz, F=73.4 MHz. .delta.F depends on the respective rotation rate and is, for example, 0.105 kHz for .OMEGA.=100 degrees/h.

The signals having the frequencies f1, f2 are produced as follows.

As mentioned above, the output of the optical/electrical transducer 8 of the evaluating unit 80 provides a square-wave signal whose amplitude is proportional to the rotation rate .OMEGA., and whose frequency is F.sub.S (switching frequency between f1 and f2). This signal is passed through an amplifier 9 and a band-pass filter 10 of midfrequency F.sub.S. The output signal of the band-pass filter 10 is applied to a first device 11, in which the sign of the rotation rate is determined, and to a second device 12, which produces a control signal proportional to the magnitude of the rotation rate.

The device 11 contains a limiter 111 followed by an AND gate 112. The second input signal of the AND gate 112 is the switching signal F.sub.S (FIG. 2b), which serves to switch between the modulation frequencies f1 and f2. The output signal of the AND gate is applied to a threshold integrator 113. The threshold is chosen to be exceeded in case of clockwise rotation and not to be exceeded in case of counterclockwise rotation, i.e., the integrator provides a 0 signal or a 1 signal depending on the direction of rotation.

The signal F.sub.S applied to the threshold integrator 113 direct. It resets the threshold circuit to zero after every period of F.sub.S.

The 0 or 1 output signal of the threshold integrator 113 is applied to an indicating device 13, where it controls the indication of direction. It is also fed to the second device 12.

The device 12 contains a half-wave rectifier 121, a mixer 122, an integrator 123, and a voltage-controlled crystal oscillator 124 in series. The second input of the mixer 122 is fed with the output (0 or 1) from the integrator 113. Depending on whether the signal is 0 or 1, the sign of the amplitude of the half-wave signal provided by the half-wave rectifier 121 will be reversed or not reversed. Accordingly, the output signal of the integrator 123 has a positive or a negative sign. Depending on the sign of the control signal for the VCXO, the output frequency F* of the latter will be raised or lowered by .delta.F (.delta.F.about..OMEGA.). The signal having the frequency F*.+-..delta.F controls the indication of the rotation rate in the indicating device 13. To this end, the indicating device incorporates a frequency meter (e.g., a counter). From the frequency measured in the evaluating unit, the constant frequency F* must be deducted. This will be illustrated by a numerical example. In the evaluating unit 80, the frequency F*.+-..delta.F is measured. Since F*=249,547 MHz, .delta.F=0.105 kHz, Thus, ##EQU5## To achieve the switchover mentioned above, the output signal of the VCXO 124 is fed to a modulating-signal-generating circuit 14, where it passes through a band-pass filter 142 and is applied to an amplifier 145. An oscillator 141 generates a signal at frequency F, which, after being mixed with the VCXO output signal F*.+-..delta.F in a mixer 41, passes through a bandpass filter 143 and is applied to an amplifier 146. A switch 147 switches at the frequency F.sub.S between the signals having the frequencies f1=F*.+-..delta.F and f2=F*.+-..delta.F+F. These signals are the modulating signals for the second Bragg cell 6.

The individual components and their operation are familiar to those skilled in the art and, therefore, will not be explained here.

So far it has been assumed that only one second Bragg cell 6 is present. It is also possible, however, to pass the two component beams through one Bragg cell each before launching them into the optical waveguide. The modulation frequencies must then be chosen in accordance with the above considerations.

To reduce external influences on the optical waveguide to a minimum, the waveguide is advantageously embedded in a sealing compound. The sealing compound must be so constituted that sound waves incident from outside are absorbed, that as uniform a temperature distribution as possible results along the optical waveguide, and that its optical refractive index is approximately equal to the optical refractive index of the cladding of the optical waveguide. A suitable sealing compound is the material used for the secondary coating.

As mentioned above, during the frequency modulation in a Bragg cell, the modulated beam is slightly deflected as a function of the modulation frequency. This is utilized in another development to provide a triaxial measuring setup which requires only one evaluating unit.

The coiled optical waveguide 7 is replaced with three coiled optical waveguides.

The axes of these coils are mutually at right angles. With this arrangement, rotation rates about three directions can be measured.

The modulation frequencies for the three optical waveguides are chosen so that the component beams traverse only one optical waveguide at a time. The modulation frequencies thus depend, on the one hand, on the above phase conditions, and, on the other hand, on the spatial arrangement of the optical waveguides in relation to the Bragg cell. At a modulation frequency of 250 kHz, the angle between the unmodulated beam and the modulated beam is 2.4 degrees.

The rotation rate is evaluated using time-division multiplexing. Only one evaluating unit is required.

The result of the measurement is affected by noise. The noise is caused mainly by the interference of the light scattered back into the fiber with the component beams travelling along the optical waveguide. It can be reduced by using lasers with short coherence lengths. However, this has various disadvantages.

A reduction of the coherence length in lasers having a long coherence length is advantageously achieved by modulating the light beam in a suitable manner. A suitable modulation is, for example, a frequency modulation with a sinusoidal signal or a statistical modulation.

The modulating signal is preferably applied to the first Bragg cell 2. It may also be applied to two second Bragg cells 6. In the latter case, however, care must be taken to ensure that this modulation causes no additional variations in the difference between the frequencies of the output signals of the second Bragg cells.

A further development makes it possible to provide control which holds constant the position of the operating point (FIG. 2a) around which .phi..sub.S varies.

As in the above-described case of rotation-rate measurements, two modulating signals with different frequencies f1 and f3 are applied to the Bragg cell 6 (FIG. 1) in succession. Now, however, the frequencies are chosen so that, after traversal of the optical waveguide, the two component beams differ in phase by 2.pi. or an integral multiple thereof (if the rotation rate is zero). The frequencies necessary for this are again calculated by the equation ##EQU6## The modulating frequencies can be fixed and kept highly stable with crystal accuracy.

As can be seen from the diagram of FIG. 2a, the output signal has the same intensity after every cycle of 2.pi.. This is also the case if rotation is taking place, i.e., if the phase shift between the two component beams is

Switching between the two modulation frequencies f1, f3 is effected at the frequency F.sub.S.

The evaluation corresponds to the evaluation of the rotation rate .OMEGA..

To make it easier to understand how the corrections are performed, the phase differences are first assumed to be .pi./2 (produced with the modulating f1), 3.pi./2 (produced with the modulating signal f2), and 5.pi./2 (produced with f3). If n or L in the equation for .phi..sub.B (page 6) varies, the position of the operating point changes as well. As a result, the phase differences between the signals obtained by modulation with the unchanged signals f1, f2 and f3 are no longer .pi. and 2.pi..

This is utilized to change f3 in such a way that the phase difference becomes 2.pi. again. To do this, the frequency of the input signal of the VCXO 124 is changed by .+-.F.sub.E, i.e., f1=F.sup.+.+-.F.sub.E. This correction applies for the phase value at .phi..sub.B1. Since, if the above quantities are assumed to vary, the period of the curve for I.sub.1 has a value different from that in the unperturbed condition, corrections for the frequencies f2 and f3 are not equal to .+-.F.sub.E, but it is necessary to take into account the respective phase value, i.e., to take care that the correction is performed at a suitable point on the curve I.sub.1. For the assumed phase differences .phi..sub.B1 =.pi./2, .phi..sub.B2 =3.pi./2, and .phi..sub.B3 =5.pi./2, the modulation frequencies change as follows (if .OMEGA.=0):

If .OMEGA..noteq.0, the modulation frequencies f1, f2 and f3 are

If the operat1ng points are not at .pi./2, 3.pi./2, and 5.pi./2 as shown in FIG. 2a, but, as is the case with modulation frequencies of about 249.547 MHz, about 249.621 MHz, and about 249.694 MHz, at phases of .phi..sub.B1 =6,801.multidot..pi./2, .phi..sub.B2 =6.803.multidot..pi./2, and .phi..sub.B3 =6,805 .pi./2, the modulation frequencies (for .OMEGA.=0) are

The correcting quantity F.sub.E is a measure of changes of n and L. Assuming that n is constant, and that R and L undergo equal relative changes, the changes R can be estimated and taken into account for the scale factor. This makes it possible to allow for variations of L and R in the evaluation of .OMEGA..

In the simplified case (.pi./2, 3.pi./2, 5.pi./2), ##EQU7## When the correction is to be made, the evaluating unit is no longer driven by .delta.F direct, but a computer in the evaluating unit 13 is fed with .delta.F in digital form, from which it calculates the rotation rate .OMEGA. by the equation ##EQU8## R was previously calculated by this computer according to the equation ##EQU9## F.sub.E was previously determined and is also fed to the computer in digital form. The other quantities for calculating .OMEGA. are constant and are stored in the computer. This computer then controls the indication of the magnitude of .OMEGA..

How such a calibration facility may be implemented will now be explained with the aid of Fig. 3.

In the embodiment of Fig. 1, a signal with the frequency F*.+-..delta.F was present at the output of the VCXO 124. In the embodiment of Fig. 3, the operations "measurement" and "calibration" are assumed to take place successively, not simultaneously. During measurement, switching is effected at the frequency F.sub.S between the modulating signals f1 and f2; during calibration, switching is effected at the frequency F.sub.S between the modulating signals f1 and f3. A controller 311 determines whether a switch 309 or a switch 310 transfers the signals for measurement or calibration, respectively. The calibration is preferably performed while .OMEGA. is changing only little. The control may also be initiated by manual operation.

The output signals of the VCXO have the frequency F*.+-..delta.F during measurement and the frequency F*.+-.F.sub.E during calibration. A frequency meter 300 measures the respective frequency and feeds it to a computer 302 through an A/D converter 301. In the computer 302, which also holds the frequency F*, control signals are produced from F*, F, .delta.F and F.sub.E which cause frequency synthesizers 306, 307 and 308 to provide signals having the frequencies f1, f2 and f3.

The signals f1 and f2 are applied to the switch 309, and the signals f1 and f3 are applied to the switch 310. The three frequency synthesizers may be replaced with a single programmable frequency synthesizer.

If calibration and measurement are to be performed simultaneously, some of the devices must be duplicated, and switchover must be effected between the calibration and measurement modulating signals at different frequencies which have a given relationship to each other.