GB2134248A - Closed loop fibre-optic gyroscope - Google Patents

Abstract

A fibre-optic interferometer gyroscope includes means 16, 17 for applying a first phase modulation in the loop operated path at a frequency 1 DIVIDED 4 tau , where tau is the loop transit time, and means 18, 19 for synchronously applying a second phase nulling modulation at a frequency 1 DIVIDED 2 tau . In synchronism with the phase nulling modulation the output of the loop is sampled. The phase nulling modulation is controlled by a feedback signal from the gyroscope output. The gyroscope is thus operated in a phase nulled mode with the Sagnac signal cancelled. <IMAGE>

Description

SPECIFICATION Closed loop fibre-optic gyroscope This invention relates to fibre-optic interferometer gyroscopes utilising the Sagnac effect.

The use of a multi-turn coil of optical fibre in which, by means of beam splitters and combiners, light from a single laser is propagated in both directions simultaneously to provide rotation sensitive output signals at a photodetector is known. Such an arrangement is described in, for example, "Sensitivity analysis of the Sagnac-effect optical-fibre ring interferometer" by Shih-Chun Lin and Thomas G. Giallorenzi in Applied Optics, Vol. 18, No. 6, 15 March, 1979.

When the output signals are combined interference fringe patterns are developed which, in a stationary system, form a fixed pattern whose shape depends on the nature of the imaging optics. If the system is rotated about the coil axis fringe excursions take place and by suitable processing rotational rate information can be extracted.

Depending on the physical details of the system (e.g operating wavelength A, fibre length L etc.) and the range of rotational velocities to be monitored, systems operating within a single fringe or over many fringes can be envisaged. Considering initially operation within one fringe measurement an examination of the form of the output signal will show that there are measurement difficulties, three of which are (i) the static nature of the sensor output (d.c. for constant angular velocity), (ii) the non-linearity of output current with phase deviation, and (iii) the pedestal level arising from spurious optical signals.

The basic homodyne system can be improved, as shown by Lin and Giallorenzi, by taking a second matching complementary fringe pattern and applying this to a differential amplifier. The differential phase shift of 180 in the second fringe can be conveniently arranged by providing an extra reflection in its transmission path. This technique eliminates common mode signals, particularly the average level of the noise (i.e. the pedestal signal). A further benefit of this arrangement is that the two outputs can give a measure of the total energy at the sensor output and a feedback control signal can be derived for maintaining constant source power.

It is difficult however to get adequate stability in such d.c. systems and, in addition, low frequency noise can be serious with some detectors. A translation of the measurement to an intermediate frequency eliminates these probiems. Lin and Giallorenzi suggest some principles for effecting modulation and synchronous detection in a homodyne system. However such modulation is performed it is clear that the counter-rotating optical signals must be separated and a differential phase perturbation applied.

According to the present invention there is provided a fibre-optic interferometer gyroscope including means for applying a first phase modulation in the loop optical path at a frequency with period 2T, where T is the transit time in the optical loop, means for applying a second phase modulation in the loop at a frequency with period 4T, and means for sampling the loop output at a frequency with period 2T in synchronisation with said first phase modulation periods which in turn is synchronised with said second phase modulation periods, and feedback control means to which the output of the gyroscope is applied, said feedback control means producing phase locked control signals for the phase modulation means.

The invention also provides a method of operating a fibre-optic interferometer gyroscope asymmetrically in the time domain wherein a first phase modulation at a frequency f = where T is the optical transit time in the fibre-optic loop, is applied to the optical signals in the loop synchronously with both a second phase modulation at a frequency f = iT and a sampling of the loop output for the gyroscope at a frequency f = +x, the amplitude and sense of the modulation at frequency f = > being controlled by a feedback signal derived from the interferometer output and being a measure of the rotation rate.

Embodiments of the invention will now be described with reference to the accompanying drawings, in which: Figure 1 illustrates schematically a closed loop fibre-optic gyroscope according to the invention, Figure 2 illustrates the phase deviation output characterisation from a basic fibre-optic gyroscope arrangement, Figure 3 illustrates driving waveforms and for the arrangement of Fig. 1, Figure 4 illustrates a form of integrated optics phase nuller and modulator for use with the arrangement of Fig. 1, and Figures 5-7 illustrate phase modulation waveforms.

The fibre-optic gyroscope shown in Fig. 1 consists essentially of a single or multi-turn coil of optical fibre 10, which is coupled via focussing lenses 11, 1 2 and a balanced beam splitter 1 3 to a laser 14 and a photodetector 1 5. (ignore for the moment the other components in the Figure). Light launched from the laser 4 is split equally at the beam splitter 1 3 and coupled into each end of the fibre 10, where it is propagated round the coil in both directions simultaneously. Upon emergence the two light outputs from the fibre are each split again equally at the beam splitter and half of each output will reach the photodetector 1 5. The two half outputs reaching the photodetector will mutually interfere at the plane of the photodetector.In general the superposition of the two output waves results in an interference pattern of concentric interference rings. In a well adjusted optical system only the central fringe is present and this central area is focussed onto the photodetector. If now the gyroscope is rotated about the axis of the coil, phase differences occur in the two outputs from the fibre which give rise to a change of light intensity at the photodetector. The photodetector response to the changing phase deviation AS arising from the rotation will have the form shown in Fig. 2, in which the output current i is at a central peak for zero rotational velocity falling to a first null and then rising to a second peak and so on as the speed of rotation is increased.

To eliminate the inherent d.c. nature of the output when the gyroscope is rotated at a constant angular velocity, phase modulation of the optical signals can be utilised. To illustrate how this phase modulation is accomplished consider a phase shifter 16, of electro-optic or other type, positioned at one end of the fibre loop or coil as in Fig. 1. This phase shifter is driven by a phase modulator 17 which applies a bias signal to the shifter at a frequency ir, where r is the optical transit time in the loop or coil 10. As a consequence of the asymmetric placement of the phase shifter 16 the clockwise and counterclockwise waves will experience different electrically derived increments of phase shift.This leads to a phase modulation on the Sagnac signal at frequency ir with resulting amplitude modulation at the photodetector output of the interferometer.

With the addition of a second phase modulator and synchronous sampling of the loop output, it is possible to detect a non-cyclic value of phase shift. This forms the basis for cancelling the Sagnac signal and operating the sensor in a closed loop phase nulled circuit. This mode of operation is the one actually illustrated schematically in Fig. 1 with the phase shifter 1 8 forming the phase nuller in conjunction with the synchronously sampled photodetector 1 5. Modulation is now applied in a second electro-optic phase shifter 1 8 which is also interposed between lens 11 and the fibre end and is driven by a phase nuller 1 2 which applies a bias signal at a frequency IT. As a consequence of the asymmetric placement of the phase shifter 1 8 and the frequency of the driving waveform, as indicated in Fig. 3, the clockwise and counterclockwise waves will experience an equal and opposite electrically derived increment of phase shift.This leads to a phase modulation of the Sagnac signal at the phase nulling frequency IT with resulting amplitude modulation at the photodetector output of the interferometer. With no Sagnac phase displacement, and zero modulation signal applied to the phase shifter 18, there will be zero modulation component at frequency ir at the photodetector output. This situation corresponds to phase modulation about the zero phase difference point of the curve in Fig. 2. Conversely the photodetector output at frequency +T will be at its maximum value when the Sagnac phase deviation has increased to /2. The action of the closed control loop is, via the amplitude and sense of the drive to phase shifter 18, to drive the phase difference to zero.The amplitude and sense of the drive to phase shifter 1 8 in this phase locked condition then represents a measure of the rotation rate.

It is convenient to fabricate the two electro-optic phase shifters 1 6 and 18 for phase modulation and phase nulling respectively as an integrated optical device as shown in Fig. 4.

Phase nuller 1 9 also controls the sampling of the photodetector 1 5. Driving waveforms for the two phase shifters and the photodetectors sampling together with the associated phase modulation response signals are given in Fig. 3. Separate phase modulators are preferable in order to avoid the filtering problem which would arise if both modulating waveforms were applied to the same modulator. A separate modulator for the bias signal also allows it to be made longer for handling the multi-fringe condition at the upper end of the dynamic range.

A clear advantage of the phase modulator architecture is that, by virtue of its relative simplicity, it forms an ideal circuit for implementation in integrated optics. In fact the whole of the optics, outside the fibre coil, laser, and detector, can be combined on a single four-port optical integrated circuit. As regards the system considerations, the contra-propagating waves are at precisely the same frequency and the minimum phase (path) differences which establish the phase modulation and nulling signals are, in a monolithically integrated component, under the highest degree of control. As shown in Fig. 4 the integrated optical device comprises a body 20 of lithium niobate having an optical guiding channel 21 diffused into one surface, e.g.

diffused titanium. Two sets of metal electrodes, 22, 22a for the phase nuller and 23, 23a for the phase modulator, are put down on the surface of the body adjacent and parallel with the channel 21. When a set of electrodes is biased by a voltage an electric field is set up across the channel and this field effectively alters the refractive index of the channel, causing a phase shift to be imposed on light passing along the channel.

Before considering the modulation waveforms, it is instructive to examine the conditions under which closed loop operation can be performed. If the phase difference between the clockwise (CW) and counterclockwise (CCW) beams is modulated at frequency f, then the optical output amplitude varies at 2f, providing the phase modulation is symmetrical about zero (see Fig. 5a). Asymmetric phase modulation produces an output containing components at both f and 2f (see Fig. 6a). Note also the alternative waveforms of Figs 5b and 6b. Since the Sagnac effect produces the change shown between Figs. 5 and 6 (i.e. a DC phase offset), any signal which nulls this offset (i.e. results in the situation shown in Fig. 5, even in the presence of rotation) is a measure of the Sagnac effect.This null condition can be established by detecting a lack of output at frequency f.

The fundamental difference between the phase difference modulation waveforms of Figs. 5 and 6 is that equal phase excursion about zero produces the nulled situation (Fig. 5), whereas unequal excursion do not (Fig. 6). With this consideration in mind, it is clear that the modulating waveforms can be sinewaves. Unfortunately, the addition of symmetric sinewaves cannot produce an asymmetric resultant, and so the output must be sampled, albeit at the expense of reducing the energy received by the photodetector. The sampled output waveform of a system requiring nulling, and a system in the nulled state, are shown in Figs. 7a and 7b respectively. Note that the waveform of Fig. 7a contains a component at the fundamental (modulation) frequency, whereas the waveform of Fig. 7b does not.

Assuming the modulators are at the input end of the CW beam (as shown in Fig. 1), the phases of the waves leaving the fibre loop arc Phase of Clout = # sin (#mt) + n sin (#nt + #) + s Phase of CCWout = # sin #m (t + #) + n sin (#n(t + #) + #) - s where f = modulation amplitude n = null modulation amplitude #m = modulator frequency #n = null modulator frequency T = loop transit time O = phase difference between null and modulation waveforms s = Sagnac phase The output amplitude is determined by the phase difference between these two waves, i.e.

p.d. = CCWout - CWout # p.d. = # (sin #m (t + #) - sin #mt) + n (sin (#n(t + #) + #) - sin (nt + #)) - 2s i.e. p.d. = 2# cos (#mt + #m#). sin (#m#) 2 2 + 2n cos (nt + H + #n#). sin (#n#) 2 2 -2s (1) The sampled waveform, translated to the phase difference of equation (1) must be of the form ## + n - s (2) if nulling of the 1st harmonic is to be achieved.This can be realized by choosing the sampling frequency #s as 2#m #s = k = 0, 1, 2 (3) (2k+ 1) in order to obtain the # # term. The + n term can be satisfied by arranging # = j. #s j = 1, 2 (4) The absolute choice of frequency is somewhat arbitrary, providing #m# # m#, #n# # m# m = 0, 1, 2 2 2 2 2 in which case the f and n terms become zero. If we choose to maximize the null signal, we have from (1) #n# = (4p + 1)# p = 0, 1, 2 2 2 i.e. #m = (2k + 1) (4p + 1) # Tj 2 If #m is to be minimized, choose k = p = 0 and j large.However, large j will.increase #n (from (4)), so choose j = 1.

Thus, we have #m = # i.e. fm = 1 2# 4# #s = 2#m i.e. fs = 1 2T #n = j. #s i.e. fn = 1 2T (5) Substituting into (1) gives p.d. = #2 # cos (#mt + # ) 4 + 2n cos (2#mt + # + # ) 2 -2s (6) Since the # and n waveforms are required to peak at the same instance for maximum efficiency, we should have # = 0 The sampled phase difference is p.d. = #2 # cos N# + 2n - 2s or p.d. = # #2 # + 2n - 2s (7) where N = 0, 1, 2 is the sample number Equation (7) has the required form. If 2n # 2s, the waveform is asymmetric about zero phase difference and produces an output at the fundamental (#m) frequency. If 2n = 2s, the waveform is symmetric with no output at the fundamental and the system is nulled. The amplitude n is a direct measure of the Sagnac phase s. The optimum modulation frequency is determined by the transit time of the loop and is given by fm = 1/4#, but this can be reduced, if necessary, at the expense of having to provide a larger drive amplitude. The null modulation and sampling frequencies can both, conveniently, be set at 2fm

Claims (8)

1. A fibre-optic interferometer gyroscope including means for applying a first phase modulation in the loop optical path at a frequency with period 2r, where T is the transit time in the optical loop, means for applying a second phase modulation in the loop at a frequency with period 4r, and means for sampling the loop output at a frequency with period 2T in synchronisation with said first phase modulation periods which in turn is synchronised with said second phase modulation periods, and feedback control means to which the output of the gyroscope is applied, said feedback control means producing phase locked control signals for the phase modulation means.

2. A gyroscope according to claim 1 wherein the means for applying the first and second phase modulation comprise first and second electro-optic phase shifters coupled to the end(s) of the fibre-optic loop.

3. A gyroscope according to claim 2 wherein said electro-optic phase shifters are realised as integrated optics devices each comprising a diffused channel optical waveguide in a body of lithium niobate with biasing electrode means adjacent the channel whereby the refractive index of the channel may be modulated.

4. A gyroscope according to claim 3 wherein the channel is formed by the diffusion of titanium.

5. A gyroscope according to claim 3 or 4 wherein both the first and second phase shifters are formed as single integrated optics device having a common channel with two sets of biasing electrodes.

6. A gyroscope according to any preceding claim wherein the light source is a laser.

7. A fibre-optic interferometer gyroscope substantially as described with reference to the accompanying drawings.

8. A method of operating a fibre-optic interferometer gyroscope asymmetrically in the time domain wherein a first phase modulation at a frequency f = IT, where T is the optical transit time in the fibre-optic loop, is applied to the optical signals in the loop synchronously with both a second phase modulation at a frequency f = +T and a sampling of the loop output for the gyroscope at a frequency f = IT, the amplitude and sense of the modulation at frequency f = IT being controlled by a feedback signal derived from the interferometer output and being a measure of the rotation rate.