Compact three-frequency resonant optical gyroscope

This application claims priority to foreign French patent application No. FR 1600581, filed on Apr. 6, 2016, the disclosure of which is incorporated by reference in its entirety.

The field of the invention is that of optical gyroscopes, in particular used in the field of inertial navigation. More precisely, the field of the invention is that of passive resonant optical gyroscopes.

Optical gyroscopes are based on the underlying principle of measurement of the Sagnac effect. The latter induces, under the effect of a rotation, a path-delay difference between two electromagnetic signals propagating in opposite directions along a ring-shaped path. This path-delay difference, which is proportional to the angular velocity of the device, may be measured either as a phase difference in the case of an interferometer set-up, or as an eigenfrequency difference between two contra-rotating modes of a ring cavity.

In the first case, it is necessary to use an optical fibre to maximise the length of the interferometer and therefore the sensitivity of the device. An interferometric fibre optic gyroscope (I-FOG) is then spoken of.

In the second case, the difference between the eigenfrequencies of the modes of the cavity may be measured in two ways. The first consists in using an active cavity, i.e. one containing a gain medium, and in measuring the frequency difference between the contra-rotating modes emitted by the cavity. Ring laser gyroscopes or RLGs are then spoken of. The second way consists in using a passive resonant cavity and in probing the eigenfrequencies of the contra-rotating modes using a laser. A passive resonant gyroscope is then spoken of.

The passive resonant gyroscope configuration possesses a certain number of advantages with respect to the other configurations. With respect to the RLG configuration, it in particular avoids the need to use a gaseous gain medium and the high-voltage system of electrodes that is conventionally associated therewith. With respect to the I-FOG configuration, it has the advantage of a much shorter optical path, this decreasing sensitivity to the environment and increasing compactness. Lastly, it employs only standard components. Thus, in particular, there is no need to use a superluminescent source.

However, although these three types of gyroscopes i.e. the I-FOG, RLG and passive resonant gyroscope, have all been demonstrated experimentally, currently, only the first two have found use in industrial applications. One of the things holding back the development of passive resonant gyrometers is the problem of backscattering of the light, which creates coupling between the contra-rotating modes, this creating a nonlinearity in the frequency response and a “blind spot” just as in conventional laser gyros, and degrading the performance of the system.

One solution to the problem of coupling between contra-propagating modes is described in document FR 1302311. This system probes the eigenfrequencies of the contra-rotating modes of a ring cavity while avoiding the problems conventionally created by backscatter, and while simultaneously providing a measurement of cavity length intended to be used to evaluate (and possibly to servocontrol to a constant value) the scale factor of the passive resonant gyroscope thus produced.

The underlying principle of this system is to use three beams at three different frequencies (instead of the two used in conventional gyroscopes). The system includes a ring cavity and a laser the emission of which is divided into three beams of different optical frequencies. By way of example, the cavity may consist of a hollow fibre to limit the Kerr effect. Each frequency is separated from the two other frequencies by a value corresponding to an integer multiple of the free spectral range of the cavity. The free spectral range (FSR) of the cavity is conventionally:

FSR=c/L c being the speed of light and L the optical length of the ring cavity.

The first beam is servocontrolled to one mode of the cavity in one propagation direction and the two others are servocontrolled to two other modes of the cavity corresponding to the opposite propagation direction. It will be noted that it is also possible to invert the servocontrol, i.e. to servocontrol a first eigenfrequency corresponding to a first resonant mode of the cavity to the first frequency of the first optical beam, for example by servocontrolling the length of the cavity.

The frequencies of the three beams are at any given time sufficiently far apart for the effect of coupling between the beams to be negligible.

In the absence of rotation, each beam is servocontrolled to one different eigenfrequency of the cavity, denoted:

f 1 =N 1 ·c/L for the first beam;

f 2 =N 2 ·c/L for the second beam; and

f 3 =N 3 ·c/L for the third beam;

where N 1 , N 2 and N 3 are known integer numbers that are all different.

The frequencies must be sufficiently close for the difference between the frequencies of each pair of beams to be compatible with the passband of a photodiode.

In the presence of a rotation, the frequency difference of the two beams propagating in the same direction allows the length of the cavity to be determined, whereas the frequency difference between two contra-rotating beams combined with the information on the length of the cavity allows the angular velocity of the assembly to be determined.

Thus, a gyroscope operating with 3 frequencies comprises means for measuring the frequency difference of the two beams propagating in the same direction, and for measuring the frequency difference between two contra-rotating beams, these two frequency differences combined together allowing the length of the cavity and the angular velocity of the cavity about an axis perpendicular to the cavity to be determined. To simplify the description, these conventional measuring means are not shown in the figures.

Specifically, in the presence of a rotation, the eigenfrequencies are shifted by an amount Ω proportional to the angular velocity, thereby giving:

f 1 =N 1 ·c/L+Ω/ 2;

f 2 =N 2 ·c/L−Ω/ 2; and

f 3 =N 3 ·c/L−Ω/ 2.

At any given time, the length of the cavity may be determined by measuring the frequency difference Δfp between the beams propagating in the same direction, i.e. in the above example Δf 2-3 :

The speed of rotation is deduced therefrom by measuring the frequency difference Δfp between two beams propagating in opposite directions, in the above example Δf 1-2 :

The architecture proposed in document FR 1302311 is shown in FIG. 1 with mirrors. The solid lines correspond to optical paths and the dashed lines to electrical connections. The laser L emits a beam that is divided into three beams F′ 1 , F′ 2 and F′ 3 . To simplify the description, the means for measuring Δfp and Δf 1-2 have not been shown.

F′ 1 is for example injected into the ring optical cavity C of length L in the counterclockwise or CCW direction, whereas the two beams F′ 2 and F′ 3 are injected into the cavity in the clockwise or CW direction. The portion of the beams F′ 2 and F′ 3 transmitted through the coupler 10 (semi-silvered mirror) passes through the optical fibre and is reflected by the optical coupler 11 , then the coupler 10 , so as to form the cavity. The portion of the beam F′ 1 transmitted through the coupler 11 passes through the optical fibre and the coupler 11 , and is reflected by the coupler 10 so as to form the cavity.

At resonance, the back-reflected intensity output from the cavity is minimal, and this property is used to servocontrol the frequencies of the three beams to the eigenmodes of the cavity. For example, the beam 101 reflected by the coupler 11 downwards in FIG. 1 is used to servocontrol the frequency of F′ 1 . Said beam corresponds to the coherent superposition of the portion of the beam F′ 1 directly reflected by 11 and the portion formed by the beams propagating in the cavity in the CCW direction, which result from the superposition of beams having made one, two, three, etc. complete circuits of the cavity in the CCW direction. Likewise, the beams 102 and 103 directed upwards in FIG. 1 are respectively used to servocontrol the frequencies of F′ 2 and F′ 3 . Said beams correspond to the coherent superposition of the portion of F′ 2 and F′ 3 directly reflected by the coupler 10 and the portions of F′ 2 and F′ 3 transmitted by this coupler 10 , then guided through the cavity in the CW direction, then reflected by 11 and lastly transmitted by 10 , corresponding to the superposition of the beams having made one, two, three, etc. complete circuits of the cavity in the CW direction.

The beam F′ 1 is servocontrolled to an eigenmode of the cavity by controlling the laser L directly using the photodiode PhD 1 and a servocontrol device DA′ 1 , which includes an optical portion DA′o 1 and an electrical portion DA′e 1 .

The beams F′ 2 and F′ 3 are servocontrolled to eigenmodes of the cavity using the photodiode PhD 23 and servocontrol devices DA′ 2 , DA′ 3 , which each include an optical portion (DA′o 2 , DA′o 3 ) acting directly on the optical frequency, and an electrical portion (DA′e 2 , DA′e 3 ).

More generally one of the beams has an eigenfrequency maintained at resonance—in the nonlimiting example, the beam F′ 1 (but it could be one of the other beams)—by direct servocontrol of the laser in the variant Opt 1 illustrated in FIG. 2 . According to another variant Opt 2 illustrated in FIG. 3 , the frequency of the beam F′ 1 is maintained at resonance by directly servocontrolling the length L of the cavity, for example using a piezoelectric modulator.

We will now explain the way in which the laser is directly servocontrolled, such as illustrated in FIG. 2 . The beam F′ 1 passes through a phase modulator PM 1 so as to generate sidebands required to obtain a frequency error signal ε 1 allowing the frequency to be servocontrolled so as to place (absence of rotation) or keep (presence of rotation) the frequency of the beam F′ 1 in resonance with the cavity mode in question. This method is based on the technique called the Pound-Drever-Hall technique, named after its inventors, and which is well known to those skilled in the art.

The beam 101 is modulated by a phase modulator PM 1 , which is placed in the optical portion DA′o 1 , so as to create sidebands at frequencies separated from the initial frequency f′ 1 by multiples of the modulation frequency, f m1 , applied by the oscillator Os 1 via PM 1 .

This frequency is chosen, if possible, to be higher than the width of the resonance of the cavity (and lower than the free spectral range of the cavity) so that the sidebands are not in resonance with the cavity. To simplify the explanation, only the two first sidebands, separated by ±f m1 from the initial frequency f′ 1 , will be considered. The beam 101 (which therefore has three spectral components at f′ 1 −f m1 , f′ 1 and f′ 1 +f m1 ) is detected by a photodiode PhD 1 , the output signal of which is demodulated by the modulating signal applied to PM 1 with an adjustment of their respective phases (phase shifter Dph 1 ) requiring the use of an electrical mixer M 1 . A lowpass filter (not shown) then allows the DC component of the demodulated signal to be isolated, the amplitude ε 1 of which is then proportional to the difference between the frequency f′ 1 of the laser and the resonant frequency of the cavity. Specifically, when the frequency f′ 1 of the laser and the resonant frequency of the cavity differ slightly, the two sidebands are unchanged (if they are far out of resonance) whereas the phase and amplitude of the beam at the frequency f′ 1 varies (since it is no longer in resonance). The coherence properties of the three spectral components of 101 then allow these fluctuations (three-beam interference) to be measured as they cause a linear variation in the demodulated signal which may thus be used as a frequency error signal, ε 1 cancelling out when the beam F′ 1 is resonant with a mode of the cavity. Servocontrol is then achieved with this signal, via control electronics ER 1 , using a conventional servocontrol method, for example, non-limitingly, PI or PID control devices (PID standing for proportional integral derivative, a reference to the three modes of action on the error signal of the control electronics). This type of control, which allows the error signal to converge on a zero value, is well known in automatic control.

Regarding the choice of the modulation frequency to be applied to PM 1 , if the finesse of the cavity is high, the width of the cavity will be small relative to the free spectral range and it will be possible to choose a modulation frequency that is very high relative to the frequency width of the resonant peaks of the cavity. The optimal situation, corresponding to the preceding explanation, will then be achieved for this servocontrol. In contrast, if the finesse of the cavity is not very high, the modulation frequency will be close to the frequency width of the resonant peaks of the cavity. The sidebands will then be partially modified when the frequency f′ 1 differs from the resonance and the servocontrol will be less effective.

The servocontrol loop controls the laser for example via the current injected ( FIG. 2 ) so as to set (absence of rotation) or keep (presence of angular rotation Ω) the frequency f′ 1 of the laser equal to a resonant frequency of the cavity:

f′ 1 =N 1 ·c/L+Ω/ 2

In the embodiment in FIG. 3 , it is the length of the cavity that is servocontrolled, the frequency of the laser remaining constant.

Thus, the optical portion DA′o 1 of the servocontrol device DA′ 1 comprises the phase modulator PM 1 and the electrical portion DA′e 1 connected to the output of the photodetector PhD 1 comprises a demodulating portion containing the phase shifter PhD 1 , the mixer M 1 , and the oscillator Os 1 of frequency f m1 , which also supplies PM 1 ; and the control electronics ER 1 .

An exemplary servocontrol mechanism for controlling the frequencies f 2 and f 3 respectively of the beams F′ 2 and F′ 3 is schematically shown in FIG. 4 . The servocontrol device is the same as for F′ 1 . However, as there is only a single laser (or only a single cavity) it is no longer possible to act on these elements. It is therefore necessary to introduce two additional components to achieve two additional degrees of freedom allowing f 2 and f 3 to be servocontrolled.

Thus, the beam F′ 2 passes through an acousto-optical modulator AOM 2 intended to modify the frequency thereof (alternatively a phase modulator allowing frequency changes to be achieved by serrodyne modulation may be used), then the transmitted portion is injected into the cavity in the CW propagation direction.

In the absence of rotation, the average value of the frequency offset, denoted Δfa in FIG. 4 , is chosen equal to a multiple of the free spectral range FSR. To this average value is also added (via the AOM 2 ) a modulation signal intended to generate the sidebands required to obtain the signal allowing this average value to be servocontrolled so as to place (absence of rotation) or keep (presence of rotation) the frequency of the beam F′ 2 in resonance with the cavity mode in question. The frequency f′ 2 of the beam F′ 2 is then servocontrolled via Δfa to an eigenmode of the cavity the frequency of which is offset from the frequency f′ 1 by a chosen amount, and respects, taking into account a possible rotation at the angular velocity Ω:

f′ 2=( N 1 +1)· c/L+Ω/ 2 i.e. Δ fa=c/L−Ω/ 2

To do this, the beam 102 described above is detected by a photodiode PhD 23 (which is the same for both beams F′ 2 and F′ 3 ). It is then processed in the same way as explained for the beam F′ 1 with the same considerations regarding the choice of the frequency of the local oscillator Osc 2 (frequency f m2 ) that modulates AOM 2 and serves in the demodulation phase.

An error signal ε 2 is thus generated that cancels out when the beam F′ 2 is resonating with the mode of the cavity.

The procedure is the same with F′ 3 except that the frequency of the oscillator Osc 3 is different from that of the oscillator Osc 2 but must meet the same criteria as F′ 1 and F′ 2 with respect to the frequency width of the resonant peaks of the cavity and its free spectral range. It is thus possible from the single signal delivered by the photodiode PhD 23 to generate two distinct frequency error signals, ε 2 and ε 3 , for F′ 2 and F′ 3 , respectively.

This signal ε 2 is used by the control electronics ER 2 , for example a PID controller, to control the acousto-optical modulator AOM 2 so as to maintain the frequency f′ 2 of the beam F′ 2 at resonance with the mode of the cavity. To do this, the aforementioned modulation signal is obtained via the adder S 2 and the oscillator Os 2 , thereby generating sidebands allowing the modulated signal that is detected by the photodiode to be obtained.

Thus, the optical portion DA′o 2 of the servocontrol device DA′ 2 comprises the acousto-optical modulator AOM 2 and the electrical portion DA′e 2 connected to the output of the photodetector PhD 23 comprises the phase shifter Dph 2 , the mixer M 2 , the oscillator Os 2 , the adder S 2 , and the control electronics ER 2 .

In this system, the acousto-optical modulators are used both to servocontrol the frequency of the corresponding beam (f′ 2 or f′ 3 ) to an eigenfrequency of the cavity, which frequency is different from the frequency f′ 1 and offset from f′ 1 by a chosen amount (corresponding to an integer multiple of the FSR that is different for each beam), and to track in real-time the shift in this eigenfrequency due to the rotation Ω. The acousto-optical modulator must therefore be able to create a frequency offset of at least one free spectral range FSR (at least for example N2=N1+1 and N3=N1+2), this introducing a limit on the minimum length of the cavity.

An AOM is typically limited to an offset of about 1 GHz, namely a cavity of a length of 20 cm (taking in this example a cavity made of optical fibre the refractive index of which is 1.5). A cavity of a clearly smaller length would therefore no longer be compatible. In addition, AOMs are bulky and consume a lot of power (typically the RF powers may be of the order of several watts). To produce a sufficiently long cavity, while maintaining, for reasons of bulk, a diameter of a few centimeters to a few tens of centimeters, the optical fibre of the cavity is looped a number of times.

Thus, the presence of acousto-optical modulators in the 3-frequency system of document FR 1302311 makes it incompatible with a “short” cavity.

However a short-cavity optical gyroscope would have a number of advantages:

decrease in the temperature sensitivity of the fibre achieved by decreasing the number of loops;

compatibility with free-space mirror cavities (a single loop), this having the advantage of suppressing the Kerr effect which is a non-linear effect known to limit the precision of fibre gyroscopes whether they be resonant or not;

compatibility with integrated optics, this technology currently being limited to a single loop. Specifically, to produce a number of loops, it would be necessary to produce lossless crossings or indeed a nonplanar integrated optical circuit so that the path allowing the cavity to be closed could pass below or above to avoid crossings. It will also be noted that acousto-optical modulators are at the present time difficult to produce in integrated-optics technology.

One aim of the present invention is to mitigate the aforementioned drawbacks by providing a 3-frequency resonant gyroscope compatible with a short cavity and/or compatible with optical functionalities produced in integrated-optics technology.

The subject of the present invention is a passive resonant optical gyroscope comprising a cavity and operating with three frequencies, and comprising:

a first injecting laser configured to inject a first optical beam into the cavity in a first direction;

a second injecting laser configured to inject a second optical beam into the cavity in a direction opposite the first direction;

a third injecting laser configured to inject a third optical beam into the cavity in one of the two aforementioned directions,

one laser amongst one of the injecting lasers, which is selected as master laser, having a master frequency, the two other injecting lasers, which are respectively denoted the first and second slave lasers, respectively having a first and a second slave frequency;

a master servocontrol device configured to directly servocontrol the master frequency to an eigenfrequency corresponding to a resonant mode of the cavity or to servocontrol an eigenfrequency corresponding to a resonant mode of the cavity to the master frequency of the master laser;

a first servocontrol stage comprising a first and a second slave device that are configured to respectively generate a first frequency and a second frequency error signal having a minimum absolute value when the first and second slave frequencies each correspond to a resonant mode of the cavity, respectively; and

a second servocontrol stage comprising a first optical phase-locking device and a second optical phase-locking device respectively comprising a first and a second slave oscillator that are configured to generate a first and a second radiofrequency offset signal,

said first and a second optical phase-locking device being configured to respectively make coherent the first slave laser with the master laser and the second slave laser with the master laser and to servocontrol the first and the second slave frequency to resonant modes of the cavity that are different from the resonant mode corresponding to the master frequency,

each radiofrequency offset signal of the second servocontrol stage being determined from the corresponding frequency error signal of the first servocontrol stage.

Preferably, the gyroscope according to the invention furthermore comprises:

a first photodetector configured to receive one or more optical beams obtained from the one or more optical beams injected in the first direction and at least one portion of which has made at least one pass through the cavity; and

a second photodetector configured to receive the one or more optical beams obtained from the one or more optical beams injected in the second direction and at least one portion of which has made at least one pass through the cavity,

said photodetectors being configured to generate three electric signals from the three received optical beams, each electric signal being sent to the master servocontrol device or to the corresponding first or second slave device.

According to one embodiment, the master servocontrol device comprises:

a master phase modulator for modulating the optical beam of the master laser;

a master demodulating device connected to the corresponding output of the photodetector: a master phase shifter, a master oscillator operating at a preset master oscillation frequency that is also used by the master phase modulator, and a master mixer for mixing the signals output by the master oscillator and master phase shifter, a master frequency error signal being obtained from the signal output by the master mixer; and

master control electronics configured to generate a correction signal from the master frequency error signal, and to directly control the master frequency of the master laser or a length of the cavity, the frequency of the master laser remaining constant.

According to one embodiment, the master phase modulator consists of an electric signal that directly modulates the supply current of the master laser at the preset master oscillation frequency.

Advantageously, the master servocontrol device is a Pound-Drever-Hall type device.

According to one embodiment, each slave device of the first servocontrol stage comprises:

a phase modulator for modulating the optical beam of the corresponding slave laser;

a demodulating device connected to the output of the photodetector having detected the corresponding optical beam, and comprising a phase shifter, an oscillator operating at a preset oscillation frequency that is also used by the corresponding phase modulator, and a mixer for mixing the signals output by the oscillator and phase shifter, the frequency error signal being obtained from the signal output by the mixer; and

control electronics configured to generate a correction signal from the frequency error signal.

Advantageously, the phase modulator consists of an electric signal that directly modulates the supply current of the corresponding slave laser at the preset oscillation frequency.

According to one embodiment, the first and second optical phase-locking device respectively comprise a third photodetector and a fourth photodetector that is configured to respectively generate a first beat signal and a second beat signal, respectively between an optical beam output by the master laser and an optical beam output by the first slave laser, and between an optical beam output by the master laser and an optical beam output by the second slave laser.

Advantageously, each radiofrequency offset signal has a tunable reference frequency and a reference phase,

and each optical phase-locking device is configured to control the frequency of the corresponding slave laser so as to servocontrol the beat signal to the radiofrequency offset signal, the reference frequency being made equal to an integer number of free spectral ranges of the cavity using a correction signal obtained from the corresponding frequency error signal, each slave frequency then respectively being offset from the master frequency by a value corresponding to the corresponding reference frequency.

Advantageously, the integer number is such that the corresponding reference frequency is comprised in a passband of the corresponding third or fourth photodetector.

According to one embodiment, each optical phase-locking device comprises a mixer configured to convert a frequency of the beat signal into a converted frequency in the radiofrequency domain, the servocontrol being carried out on the basis of the converted frequency.

Preferably, each optical phase-locking device comprises, to achieve the servocontrol of the beat signal to the radiofrequency offset signal:

a phase comparator configured to compare respectively a phase of the beat signal or of the converted signal and the phase of the radiofrequency offset signal, the comparator being configured to generate a phase error signal; and

control electronics configured to generate a correction signal and to control the slave frequency of the slave laser on the basis of the phase error signal.

According to one variant, the first and second photodetectors are configured to receive optical beams that are at least partially reflected by the cavity.

According to another variant, the first and second photodetectors are configured to receive optical beams transmitted by the cavity.

According to one embodiment, an optical block comprising the paths of the optical beams and the optical components required to implement the gyroscope according to the invention is produced in the form of at least one photonic integrated circuit.

Other features, aims and advantages of the present invention will become apparent on reading the following detailed description and with reference to the appended drawings, which are given by way of nonlimiting examples and in which:

FIG. 1 , already cited, illustrates the architecture of a 3-frequency passive resonant gyroscope according to the prior art.

FIG. 2 , already cited, illustrates a 3-frequency passive resonant gyroscope according to the prior art with direct servocontrol of the laser corresponding to one of the frequencies.

FIG. 3 , already cited, illustrates a 3-frequency passive resonant gyroscope according to the prior art with servocontrol of the length of the cavity, the frequency of the laser remaining constant.

FIG. 4 , already cited, illustrates the servocontrol of the two other frequencies.

FIG. 5 illustrates a passive resonant optical gyroscope 50 operating with 3 frequencies according to the invention.

FIG. 6 illustrates one embodiment of the gyroscope 50 according to the invention and more particularly details the various components used.

FIG. 7 illustrates the underlying principle of servocontrol based on an optical phase-lock loop.

FIG. 8 schematically shows one embodiment of the gyroscope according to the invention the cavity of which comprises two free-space couplers and one optical fibre.

FIG. 9 schematically shows one embodiment of the gyroscope according to the invention, said gyroscope being particularly well suited when the cavity consists of an optical fibre.

FIG. 10 schematically shows one embodiment of the gyroscope according to the invention, in which the first and second photodetector are configured to receive optical beams corresponding to optical beams transmitted by the cavity.

FIG. 11 illustrates a gyroscope according to the invention comprising a photonic circuit produced in integrated-optics technology

FIG. 5 illustrates a passive resonant optical gyroscope 50 operating with 3 frequencies according to the invention. To simplify the description, the means for measuring Δfp and Δf 1-2 (see the prior art) have not been shown.

Here, the method for obtaining 3 suitable frequencies will be described.

The 3-frequency gyroscope 50 comprises a cavity C of length L. It comprises a first injecting laser L 1 configured to inject a first optical beam F 1 into the cavity in a first direction, a second injecting laser L 2 configured to inject a second optical beam F 2 into the cavity in a direction opposite the first direction, and a third injecting laser L 3 configured to inject a third optical beam F 3 into the cavity in one of the two aforementioned directions.

In the nonlimiting example in FIG. 5 , the beams F 1 and F 3 are injected in the CW direction and the beam F 2 is injected in the CCW direction.

One laser amongst one of the injecting lasers L 1 , L 2 and L 3 is chosen as master laser, in the example it is L 1 but any of the three lasers may be chosen as master laser. The master laser has a master frequency, here f 1 .

The two other injecting lasers are respectively denoted first slave laser, L 2 in the example in FIG. 5 , and second slave laser, L 3 in the example in FIG. 5 . The first slave laser has a first slave frequency f 2 in the example in FIG. 5 , and the second slave laser L 3 has a second slave frequency, f 3 in the example in FIG. 5 . In the rest of the description, the explanations are given with L 1 as master laser and L 2 and L 3 as slave lasers, but the invention applies identically for any other choice of the master laser and slave lasers.

The gyroscope 50 according to the invention also comprises a master servocontrol device DA 1 configured to directly servocontrol the master frequency f 1 to an eigenfrequency of the cavity. The servocontrol is carried out in a conventional way, as for example described in the prior art, on the basis of the frequency error signal ε 1 .

According to a first option, the error signal ε 1 is used to change the frequency of the master laser by acting on the available input (for example a modulation of the current for a semiconductor laser), the servocontrol controlling the frequency of the laser directly so that it corresponds to a resonant mode of the cavity, as illustrated in FIG. 5 (see also FIG. 2 ).

According to a second option, the error signal ε 1 is used to modulate the length of the cavity via a piezoelectric transducer (for example for a fibre laser), in order to servocontrol an eigenfrequency corresponding to a resonant mode of the cavity to the frequency of the master laser, which remains constant (see FIG. 3 ).

The gyroscope 50 according to the invention furthermore comprises a first servocontrol stage comprising a first slave device, D 2 in the example, associated with the first slave laser and a second slave device, D 3 in the example, associated with the second slave laser. The first slave device D 2 is configured to generate a first frequency error signal, ε 2 in the example of FIG. 5 , having a minimum absolute value when the first slave frequency f 2 corresponds to a resonant mode of the cavity. The second slave device D 3 is configured to generate a second frequency error signal, ε 3 in the example of FIG. 5 , having a minimum absolute value when the second slave frequency f 3 corresponds to a resonant mode of the cavity. Typically, the frequency error signals are obtained by the conventional means described in the prior art. These error signals allow the deviation of the frequency of the slave lasers from resonance to be quantified and are used by a second servocontrol stage of the gyroscope 50 according to the invention.

The second servocontrol stage comprises a first optical phase-locking device OPLL 2 - 1 comprising a first oscillator Osc 2 (associated with the first slave laser L 2 ) configured to generate a first radiofrequency offset signal, the first optical phase-locking device OPLL 2 - 1 being configured to make the first slave laser L 2 coherent with the master laser L 1 and to servocontrol the first slave frequency f 2 to a resonant mode of the cavity that is different from the resonant mode corresponding to the master frequency f 1 . Typically the radiofrequency offset signal has a frequency comprised between a few tens of MHz, for cavities based on optical fibres, up to several tens of GHz for miniature integrated-optics cavities, the maximum offset in any case being limited by the maximum passband of the detectors (typically between 40 and 100 GHz at 1.5 μm).

The second servocontrol stage also comprises a second optical phase-locking device OPLL 3 - 1 comprising a second oscillator Osc 3 (associated with the second slave laser L 3 ) configured to generate a second radiofrequency offset signal, the second optical phase-locking device OPLL 3 - 1 being configured to make the second slave laser L 3 coherent with the master laser L 1 and to servocontrol the second slave frequency f 3 to a resonant mode of the cavity that is different from the resonant mode corresponding to the master frequency f 1 . The master laser L 1 is therefore the laser directly servocontrolled to the cavity and serves as reference laser to which the two other lasers i.e. the lasers denoted slave lasers, are phase looped. The invention therefore uses two optical phase-locking devices (denoted OPLL, for optical phase-locked loop) to make the slave lasers coherent with the master laser and to make the two slave lasers lase at resonant frequencies of the cavity C.

To do this, each radiofrequency offset signal of the second servocontrol stage is determined from the corresponding frequency error signal ε 2 , ε 3 of the first servocontrol stage. In other words, for each slave laser, the gyroscope according to the invention uses the frequency error signal generated by the first servocontrol stage to control the oscillator of the corresponding phase-locked loop. The operation of an OPLL loop and particular embodiments of implementation of the two OPLL loops of the gyroscope according to the invention are described in detail below.

This architecture has the advantage, with respect to the 3-frequency architecture of the prior art, of no longer requiring the acousto-optical modulators. It will be recalled that these modulators allowed, from a single laser, three beams at different frequencies to be obtained, each however in resonance with the cavity. The gyroscope according to the invention comprises three independent lasers that are made coherent with one another while ensuring that they each have a different frequency and are in resonance with the cavity. By virtue of the use of these three independent lasers, it is possible to obtain much larger frequency differences than with acousto-optical modulators and also to improve compactness.

The coherence relationships between the beams are obtained and controlled by the OPLL. The frequency error signals ε 2 and ε 3 respectively allow the frequency of the oscillator of the OPLL 2 - 1 (between the beam F 1 and the beam F 2 ) and the frequency of the oscillator of the OPLL 3 - 1 (between the beam F 1 and the beam F 3 ) to be servocontrolled so that the first slave laser L 2 and the second slave laser L 3 are respectively in resonance.

Preferably, the gyroscope 50 according to the invention also comprises a first photodetector PhD 13 configured to receive one or more optical beams obtained from the one or more optical beams injected in the first direction and at least one portion of which has made at least one pass through the cavity, and a second photodetector PhD 2 configured to receive one or more optical beams obtained from the optical beams injected in the second direction and at least one portion of which has made at least one pass through the cavity.

These photodetectors may be placed in a number of locations with respect to the cavity depending on the type of cavity and of the type (reflected or transmitted) of optical beam being collected, as described below. In the example of FIG. 5 , the detector PhD 13 receives the beams 51 and 53 obtained from the beams F 1 and F 3 injected in the CW direction, and the detector PhD 2 receives the beam 52 obtained from the beam injected F 2 in the CCW direction.

The photodetectors PhD 13 and PhD 2 are configured to generate three electric signals from the 3 detected optical signals 51 , 52 and 53 .

The detector PhD 13 detects the two signals 51 and 53 , which are liable to beat together, but the demodulating step then carried out on each of the electric signals makes it possible to isolate only the signal of interest.

Each electric signal is sent to the corresponding device. In the example, the electric signal obtained from the optical beam 51 (F 1 ) forms the input of the master servocontrol device DA 1 ; the electrical signal obtained from the optical beam 52 (F 2 ) forms the input of the second slave device D 2 ; and the electrical signal obtained from the optical beam 53 (F 3 ) forms the input of the first slave device D 1 .

FIG. 6 illustrates one embodiment of the gyroscope 50 according to the invention and more particularly details the various components used.

Advantageously, the master servocontrol device DA 1 is of the type described in the prior art with reference to FIGS. 2 and 3 . It comprises a master phase modulator PM 1 for modulating the optical beam output by the master laser L 1 and a master demodulating device DM 1 that acts on the electrical signal output by the corresponding photodetector PhD 13 . The device DM 1 comprises a master phase shifter, a master oscillator Osc 1 operating at a preset master oscillation frequency fm 1 that is also used by the master phase modulator and a master mixer for mixing the signals output by the master oscillator and the master phase shifter, the master frequency error signal ε 1 being obtained from the signal output by the master mixer. The device DA 1 also comprises master control electronics ER 1 configured to generate a correction signal from the master frequency error signal ε 1 , and to directly control the master frequency f 1 of the master laser L 1 or to control the length L of the cavity, the frequency of the master laser remaining constant.

Advantageously, the master servocontrol device DA 1 is a Pound-Drevor-Hall type device.

According to one variant, the master phase modulator consists of an electric signal that directly modulates the supply current of the master laser L 1 at the preset master oscillation frequency fm 1 . Therefore, when for example the master injecting laser L 1 is a laser diode, a direct modulation of the current of the laser L 1 at the given frequency fm 1 is used to generate the sidebands, instead of a phase modulator component PM 1 .

FIG. 6 also describes an example of slave devices D 2 and D 3 . These devices have an architecture that is substantially identical to that of the master servocontrol device DA 1 .

Advantageously, each slave device D 2 or D 3 of the first servocontrol stage comprises:

a phase modulator PM 2 or PM 3 for modulating the optical beam F 2 or F 3 of the corresponding slave laser L 2 or L 3 .

According to one variant, the phase modulator consists of an electrical signal that directly modulates the supply current of the corresponding slave laser L 2 , L 3 , at the preset oscillation frequency fm 2 , fm 3 .

Thus, instead of the phase modulator components PM 2 and/or P 3 , to generate the sidebands, a direct modulation of the current of the lasers L 2 and L 3 at the same frequencies fm 2 , fm 3 is used, when for example the injecting lasers L 2 and L 3 are laser diodes.

Each slave device of the first servocontrol stage also comprises a demodulating device DM 2 or DM 3 connected to the output of the photodetector having detected the corresponding optical beam, 52 detected by PhD 2 or 53 detected by PhD 13 . Each demodulating device comprises a phase shifter, an oscillator operating at a preset oscillation frequency fm 2 or fm 3 that is also used by the corresponding phase modulator, and a mixer for mixing the signals output by the oscillator and phase shifter, the frequency error signal ε 2 or ε 3 being obtained from the signal output by the mixer;

control electronics ER 2 or ER 3 configured to generate a correction signal from the frequency error signal ε 2 , ε 3 .

The error signals ε 2 , ε 3 are not used to directly servocontrol the frequency of the corresponding laser L 2 , L 3 to a mode of the cavity. They are used by way of correction signal to set the frequency of the radiofrequency offset signal, as explained below.

We will now describe the second servocontrol stage, which is based on the OPLL principle.

OPLLs are based on phased-locked loops, or PLLs, and are able to modify the relative phase of two optical beams by transposition to an electrical signal. The principle of such a servocontrol is as illustrated in FIG. 7 .

It is sought to servocontrol a slave frequency fe of a slave laser Le to the master frequency fm of a master laser.

It will be recalled that the frequency of a signal f is proportional to the derivative of the phase φ of the signal with respect to time. Making the phase difference between a slave signal to be servocontrolled and a reference signal for example zero allows a slave frequency fe to be servocontrolled to the master frequency fm.

In other words, an OPLL is configured to carry out the servocontrol on the basis of an error signal ε ϕ that depends on a phase difference between on the one hand the fm−fe beat between the master frequency and the slave frequency and on the other hand a reference signal having a predefined reference frequency fref.

According to one preferred embodiment, the phased-locked loop OPLL comprises a photodiode PhD that detects the optical beams output by the master and slave laser, and more particularly the beat signal between these frequencies, of frequency fm−fe. The frequency of the slave laser fe is servocontrolled to the frequency of the master laser fm (itself directly servocontrolled to a resonant mode of the cavity of the gyroscope), on the basis of this fm−fe beat that it is desired to adjust to a preset value.

When the beat frequency between the two lasers is typically of the order of magnitude of a gigahertz, the comparison of the phases of these signals is very complex to implement. The beat frequency fm−fe is then converted into a converted signal of lower frequency fm−fe−fdc using a mixer M, this operation being referred to as a down conversion. The aim is to make the frequency of the signal to be servocontrolled compatible with the operating range of the phase comparator Compφ. The obtained signal of frequency fm−fe−fdc and of phase ϕ=2π(fm−fe−fdc)t+ϕm−ϕe (it is assumed that the oscillator at the frequency fd to use the down conversion has a sufficiently stable phase to not be taken into account) typically has a frequency comprised between 1 and 500 MHz. Preferably a filter is added after the mixer in order to select only the signal of interest of frequency fm−fe−fdc in the desired range.

The OPLL loop also comprises a reference oscillator Osc configured to generate a radiofrequency offset signal having a radio reference frequency fref and a reference phase ϕref.

Next, the phase comparator Compϕ generates an error signal ε ϕ that depends on the phase difference ϕ−ϕref between the converted signal and the radio reference signal.

Lastly, an electronic control device ER generates a correction signal and controls the slave frequency fe of the slave laser so as to minimise the error signal ε ϕ . In the case of a laser diode DFB, typically the supply current of the laser is controlled, optical frequency depending on the current.

Typically on turn on, after a certain length of time, the frequency of the converted beat signal locks to the chosen frequency fref of the oscillator, with fm−fe−fdc=+/−fref

The slave laser then has a slave frequency equal to the master frequency shifted by the reference frequency and the down conversion frequency.

Thus, when two optical beams output by two sources are made to beat on a photodetector, and their frequencies are close enough to lie within the passband of the detector, a sinusoidal beat signal is obtained the duration, frequency stability, phase stability and amplitude of which depend on the degree of coherence between the two sources. The more the sources are coherent (their intensities are assumed to remain constant), the more this beat signal resembles a sinusoidal signal generated by an electric generator, with small fluctuations in the frequency and phase (few discontinuities, jumps, or transient divergences from a sinusoid).

The objective of an OPLL is to servocontrol the beat signal between two lasers, this signal optionally being down shifted to a frequency domain accessible to phase comparators, to a reference (and therefore very stable) oscillator, the radiofrequency offset signal, so that this beat signal is a sinusoid that is as stable as possible of that delivered by the reference oscillator (the down conversion signal is assumed to be sufficiently stable to not be taken into account).

Once this regime has been achieved, the two lasers, which were precedingly independent, have a coherency relationship one with the other.

The gyroscope according to the invention uses, in an original way, two optical phase-locked loops. The coherence relationship between the three injecting lasers of the gyroscope is obtained by virtue of two optical phase-locked loops, one between the master laser and a first slave, and another between the master laser and a second slave.

Preferably, to collect the beat signals, the first optical phase-locking device OPLL 2 - 1 comprises a third photodetector PhDO 12 that receives the optical beams output by the master laser L 1 and the first slave laser L 2 and that generates a first electrical beat signal from the received optical beams. Likewise, the second optical phase-locking device OPLL 3 - 1 comprises a fourth photodetector PhDO 13 that receives the optical beams output by the master laser L 1 and the second slave laser L 3 and that generates a second electrical beat signal from the received optical beams.

In addition, if for each OPLL the reference oscillator is tunable, and one of the two lasers (the master laser) is servocontrolled to a resonance of the cavity, then it is possible to tune the frequency of the reference oscillator so that the frequency of the beat signal between the two beams (which is ideally a sinusoidal signal maintained by the OPLL) corresponds to an integer number of FSRs of the cavity. The slave laser then also lases into a resonant mode of the cavity, this mode being different from the resonant mode of the master laser.

Thus, advantageously the first radiofrequency offset signal has a tunable first reference frequency f 2 ref and a reference phase ϕ 2 ref and the first optical phase-locking device OPLL 2 - 1 is configured to control the frequency f 2 of the first slave laser L 2 so as to servocontrol the first beat signal, optionally converted to the low-frequency domain, to the first radiofrequency offset signal, the first reference frequency f 2 ref without down conversion or the sum of the first reference frequency f 2 ref and the down-conversion frequency fdc in the contrary case being made equal to an integer number n 1 of free spectral ranges FSRs of the cavity using the correction signal obtained from the first frequency error signal ε 2 delivered by the first servocontrol stage.

Likewise, the second radiofrequency offset signal has a tunable second reference frequency f 3 ref and a reference phase ϕ 3 ref and the second optical phase-locking device OPLL 3 - 1 is configured to control the frequency f 3 of the second slave laser L 3 so as to servocontrol the second beat signal, optionally converted to the low-frequency domain, to the second radiofrequency offset signal, the second reference frequency f 3 ref without down conversion or the sum of the second reference frequency f 3 ref and the down conversion frequency fdc′ in the contrary case being made equal to an integer number n 2 of free spectral ranges FSRs of the cavity using the correction signal obtained from the second frequency error signal ε 3 delivered by the first servocontrol stage.

When the frequencies are locked and in the absence of rotation, the slave frequency f 2 is offset from the master frequency f 1 by a value corresponding to the reference frequency f 2 ref+fd=n 1 ·FSR.

f 2− f 1=+/−( f 2ref+ fd )=+/− n 1·FSR.

Likewise, the slave frequency f 3 is offset from the master frequency f 1 by a value corresponding to the reference frequency f 3 ref+fd=n 2 ·FSR.

F 3− f 1=+/−( f 3ref+ fd )=+/− n 2·FSR.

When the gyroscope experiences an angular rotation, the resonant frequencies vary as a function of time. The operation of the servocontrol may be likened to that of a looped system the operating point of which is changed to ensure the resonances are tracked.

One embodiment of the OPLLs is described in FIG. 6 .

Each optical phase-locking device OPLL 2 - 1 (OPLL 3 - 1 ) comprises, to achieve the servocontrol of the beat signal to the radiofrequency offset signal:

a phase comparator PC (PC′) configured to compare a phase ϕ 2 (ϕ 3 ) of the beat signal or of the converted signal and the phase ϕ 2 ref (ϕ 3 ref) of the corresponding radiofrequency offset signal, the comparator being configured to generate a phase error signal εϕ (εϕ′);

control electronics ER (ER) configured to generate a correction signal and to control the slave frequency f 2 (f 3 ) of the slave laser L 2 (L 3 ) on the basis of the phase error signal εϕ (εϕ′).

With respect to the use of a single laser and acousto-optical modulators, a 3-laser gyroscope is obtained in which:

the coherence relationships between the beams are obtained and controlled by the OPLLs; the differences between the frequencies of the three beams, allowing the frequency of each beam to be servocontrolled so as to obtain (absence of rotation) or preserve (presence of rotation) their resonance, are controlled by controlling the frequency of the reference oscillator of each of the OPLLs; the error signal ε 1 allows as above the laser L 1 to be servocontrolled to the cavity or vice versa; the error signals ε 2 and ε 3 allow the frequency f 2 ref of the oscillator of the OPLL 1 - 2 and the frequency f 3 ref of the OPLL 1 - 3 to be respectively servocontrolled so that L 2 and L 3 are in resonance.

To ensure correct operation of the OPLLs the integer numbers n 1 , n 2 and the down-conversion frequencies fdc and fdc′ are such that the corresponding reference frequencies f 2 ref, f 3 ref are comprised in a passband of the corresponding photodetector.

According to one embodiment illustrated in FIG. 6 , a down conversion is necessary and in this case each optical frequency-locking device OPLL 1 - 2 (OPLL 1 - 3 ) comprises a mixer M (M′) configured to convert a frequency of the beat signal f 2 -f 1 (f 3 -f 1 ) into a converted frequency f 2 −f 1 −fdc (f 3 −f 1 −fdc) in the radiofrequency domain, the servocontrol being carried out on the basis of the converted frequency.

Typically each converted frequency is comprised between 1 and 500 MHz. Such a frequency is compatible with a phase comparator produced in integrated-optics technology.

According to one variant a filter is connected to the output of each mixer.

The operation of the gyroscope according to the invention may be based on various optical signals 51 , 52 , 53 received by the first and second detectors PhD 13 and PhD 2 . According to a first embodiment illustrated in FIGS. 8 and 9 , the first and second photodetector are configured to receive optical beams 51 , 52 , 53 corresponding to the optical beams at least partially reflected by the cavity. When the reflected intensities are used, the latter are maximal off resonance and minimal at resonance and the servocontrol is adapted accordingly. In the case where the gyroscope is optical fibre based ( FIG. 9 ), circulators Circ and Circ′ are inserted on the path of the optical beams in order to well separate the reflected beams from the incident beams. The use of reflected beams to deliver the error signals ε 1 , ε 2 and ε 3 is the most optimal, because in this case the maximum value of the modulation frequency of PM 1 , PM 2 and PM 3 is limited only by the free spectral range FSR.

According to a second body embodiment illustrated in FIG. 10 , the first and second photodetector are configured to receive optical beams 51 , 52 , 53 corresponding to the optical beams transmitted by the cavity. When the transmitted intensities are used, the latter are minimal off resonance and maximal at resonance and the servocontrol is adapted accordingly. One advantage is that, lasers generally comprising isolators, the presence of the isolators Iso and Iso′ is not indispensable, and the production of the cavity C and of the couplers C 1 and C 2 is compatible with an implementation in integrated optics.

The gyroscope according to the invention is also compatible with any type of cavity: free-space cavities, hollow-fibre cavities, integrated-optics resonators, etc.

According to one embodiment, the cavity C of the gyroscope 50 comprises two free-space couplers 10 , 11 , the rest of the cavity for example being in free space (mirrors) or comprising an optical fibre OF such as illustrated in FIG. 8 .

Another embodiment illustrated in FIG. 9 is particularly suitable when the cavity consists of an optical fibre OF. The cavity comprises, apart from the optical fibre OF, a first 2-by-2 coupler C 1 , this first coupler being configured to inject the optical beams F 1 , F 2 , F 3 into the cavity and to direct the optical beams 51 , 52 , 53 reflected by the cavity towards the first and second photodetector PhD 13 , PhD 2 .

According to another embodiment, the fibre-based cavity comprises a first 2-by-2 coupler C 1 and a second 2-by-2 coupler C 2 , the first coupler C 1 being configured to inject the optical beams F 1 , F 2 , F 3 into the cavity, the second coupler C 2 being configured to direct the optical beams 51 , 52 and 53 transmitted by the cavity towards the first and second photodetector PhD 13 , PhD 2 .

A nonlimiting numerical example of the gyroscope 50 is given below.

Semiconductor or fibre-optic injecting lasers L 1 , L 2 and L 3 emitting at a wavelength of 1.55 μm are for example considered. A 5 cm-long cavity with a refractive index of 1.6 is considered, i.e. a cavity with an FSR of 3.75 GHz, the cavity being produced in integrated-optics technology (in this example from Si 3 N 4 to obtain low propagation losses). This length is a good compromise because it is not necessarily being sought to make the cavity very small because sensitivity depends on area.

For a master frequency f 1 of about 1.94×10 14 Hz (λ=1.55 μm) N1 is therefore about 51000.

It is sought to offset the slave frequencies f 2 and f 3 by at least one whole FSR.

It is possible to use: +1·FSR(N2=N1+1), and −1·FSR(N3=N1−1).

It is also possible to use: +1·FSR and +2·FSR(N2=N1+1 and N3=N1+2) etc. or −1·FSR and −2·FSR(N2=N1−1 and N3=N1−2).

On account of the value of the FSR, an offset of a few FSRs must be used if compatibility with photodiodes integratable into integrated-optical circuits, etc. is to be guaranteed.

The gyroscope 50 according to the invention is compatible with integrated-optics technology, because it does not require acousto-optical modulators.

In addition, the use of OPLL loops is compatible with the orders of magnitude of the offsets to be achieved and an integrated-photonic implementation, thereby allowing bulk and total cost to be drastically decreased.

The optical block of the gyroscope 50 is defined as comprising the paths of the optical beams output by the lasers and the optical components (such as the phase modulators) necessary to implement said gyroscope and including the photodetectors.

At a first level of integration illustrated in FIG. 11 , the optical block is a photonic circuit PIC produced in integrated-optics technology, for example on a silicon or InP substrate, or on a substrate of any other material compatible with the required functionalities. Coupling into and out of the circuit PIC may be achieved via its edge face or directly out of the plane of the circuit using gratings. In addition, in FIG. 11 , each modulator PM 1 , PM 2 , PM 3 is equipped with a photodiode. This allows the power of the three beams to be measured and controlled and thus the Kerr effect associated with any difference in the power of two waves propagating in opposite directions through the cavity to be decreased. This circuit PIC may then be connected to a mirror-based cavity ( FIG. 11 ) or indeed a fibre-based cavity. Thus the optical block is produced in the form of at least one integrated circuit.

At a higher level of integration, the cavity and/or the injecting lasers are also produced in integrated-optics technology.

For the sake of clarity, the measurement of the frequencies f 1 , f 2 and f 3 allowing the length of the cavity and the speed of rotation to be determined are not shown. This measurement may be carried out in two ways.

According to one variant, a calibration of the control signals injected into the OPLLs and the master laser are used to determine f 1 , f 2 and f 3 .

According to another variant, the optical beams are used directly. For example, by sampling a portion of F 1 and F 3 before the coupler 10 ( FIGS. 5 and 11 ) or before the coupler C 1 ( FIGS. 9 and 10 ) and by transmitting this optical signal to a photodiode, f 1 -f 3 may be measured directly using a frequency counter. As it is a question of a beat, it is also possible to use RLG laser-gyro methods directly (provided that the passband is suitably adapted). Likewise, by sampling and combining F 1 , F 3 and F 2 , the electrical beat signal will have components—F 3 , F 1 -F 2 , F 2 -F 3 that may be filtered and measured. According to another variant, a configuration that is a mix of FIGS. 9 and 10 is used: the reflected beams are used for the servocontrol (this also allowing a first set of cavity-length and speed-of-rotation values to be obtained) and the transmitted beams are recombined to deduce from the beating thereof a second set of cavity-length and speed-of-rotation values.