Mean-free controlled phase modulator for fiber optic gyroscopes and fiber optic gyroscopes

Abstract

Control system (100) for a fiber optic gyroscope comprising: a phase modulator (110) for modulating a phase of a light signal (115); a control unit (120) for generating a control signal (125) by the magnitude of which the phase is modulated and which is supplied to the phase modulator (110); characterized in that the control signal (125) changes statistically and is mean-free.

Description

[0001] The invention relates to a control system for a fiber optic gyroscope, as well as a fiber optic gyroscope, in particular a fiber optic Sagnac interferometer.

[0002] The invention also relates to a control system for a fiber optic current sensor based on the Faraday effect, as well as such a fiber optic current sensor.

[0003] Fiber optic gyroscopes, such as Sagnac interferometers, are used in gyroscopes in inertial navigation systems.

[0004] Inertial navigation systems can be implemented in a variety of ways, generally based on detecting forces or accelerations acting on an object and the resulting rotation rates for position determination. Instead of mechanical effects, optical effects can also be used to determine rotation rates for inertial navigation systems. Such an inertial navigation system can be based on at least one fiber-optic gyroscope, such as a Sagnac interferometer. This utilizes the Sagnac effect, according to which an optical fiber loop rotates around its normal, resulting in an optical path difference between two light rays traveling in opposite directions within the loop.When observing the light emerging from the optical fiber loop and superimposed on the surface of the two light beams, a change in intensity becomes visible during rotation. This change can be described by an interferometer characteristic curve, which describes the intensity change as a function of the phase difference between the two light waves. In other words, a rotational movement acting on a fiber optic gyroscope, such as a Sagnac interferometer, causes a phase shift between the two oppositely rotating light beams, so that at the point of superposition of the two beams, a change in intensity dependent on the rotational movement can be observed.

[0005] The phase shift in a fiber-optic Sagnac interferometer is directly proportional to the rotational speed, the path length of the light path in the optical loop or coil, and the diameter of the circular light path. The phase shift is also inversely proportional to the wavelength of the light used.

[0006] The interferometer characteristic curve mentioned above, which describes the dependence of the light intensity, which is to serve as an observation quantity for determining the rotation, on the phase difference, is cosine-shaped.

[0007] Since a corresponding transfer function at the maximum of the cosine curve is insensitive to small input values, and the sign of the phase shift corresponding to the direction of rotation cannot be determined, phase modulation is often used to set the operating point of the fiber optic gyroscope such that it lies at the point of maximum slope of the cosine function. Sine or square wave modulations, for example, are suitable for this purpose. This ensures maximum sensitivity of the interferometer even with a small rotational movement.

[0008] A gyroscope containing a fiber optic gyroscope typically incorporates a multifunctional integrated optical chip (MIOC) as a phase modulator, enabling phase modulation of passing light beams. The MIOC is usually part of a control loop for adjusting the phase modulation described above using a control signal. Physical effects within the MIOC, such as mobile charge carriers, cause the phase modulation to depend on the frequency of the control signal. The MIOC therefore exhibits a frequency response, resulting in different frequency responses at different frequencies; in particular, the MIOC's response is less pronounced at low frequencies than at high frequencies. Since this phenomenon is also due to the mobility of the charge carriers within the MIOC, the MIOC's frequency response is also dependent on the ambient temperature.the temperature of the MIOC itself.

[0009] This MIOC frequency response leads to a specific form of lock-in effect, manifesting as a clustering of output signals corresponding to a rotation rate of 0° / h at truly low rotation rates around 0° / h. The resulting insensitivity of a fiber optic gyroscope, such as a Sagnac interferometer, to very low rotation rates hinders highly precise and reliable measurements, especially at very low rates.

[0010] The invention is based on the objective of providing a control system for a fiber-optic gyroscope in which the lock-in effect caused by the frequency response of the phase modulator is effectively suppressed and in which highly precise rotation rate measurements are possible even at low rotation rates. This objective is achieved according to the invention by the subject matter of the independent claims. Further developments are specified in the respective dependent claims.

[0011] A control system for a fiber optic gyroscope comprises a phase modulator for modulating the phase of a light signal and a control unit for generating a control signal by which the phase is modulated and which is fed to the phase modulator. The control signal changes statistically and is averaging-free.

[0012] It has been shown that by using a control signal for phase modulation in a fiber optic gyroscope, whose signal values ​​follow a statistical distribution and whose expected value is zero, the low-frequency component in the control signal can be reduced. Control signals for low rotation rates thus exhibit a similar frequency spectrum to control signals for high rotation rates. Therefore, even at low rotation rates, it is ensured that no low-frequency control signals are fed to the phase modulator for controlling the phase modulation. This makes the response behavior of the phase modulator at low rotation rates comparable to that at high rotation rates, thereby reducing the lock-in effect and eliminating any preference for the measured value "0° / h" at low rotation rates.

[0013] The control unit can generate the statistical, mean-free control signal from a deterministic primary signal. This ensures that the statistical control signal is derived from a deterministically determined primary signal. This allows the phase modulator to be set to specific groups of modulation values, even though a statistically fluctuating signal is used for direct control.

[0014] The phase modulation signal can be used, for example, to determine at which phase or phase difference the intensity of the interference signal measured in the fiber optic gyroscope should be measured; that is, the phase modulation determines the operating point for the intensity measurement on the interferometer characteristic curve. The primary signal can then be used, for example, to define an operating point. The primary signal is converted into the statistical, mean-free control signal in such a way that, instead of the operating point determined by the primary signal, other operating points are selected that yield equivalent measurement results. This ensures that, despite the statistical control signal being used to reduce the lock-in effect, measurement conditions for the intensity measurement can be determined.

[0015] The control unit can include a random number generator that produces random numbers r between 0 and q with a probability of 1 / q. For a primary signal of size x generated by the control unit, the control signal can be of the size (x – x). r ) exhibit, where x r = 0 for x + r < q and x r = q for x + r ≥ q. This ensures that the control signal, and thus the phase shift generated by the control signal, is statistically distributed and free of mean values, since the signal component x rThe mean value x, i.e., the value of the primary signal, is used. The primary signal is thus corrected (for a fixed q) by the value q, depending on the magnitude of x and the magnitude of r. This correction is statistically applied due to the random nature of r. This allows the measurement conditions to be determined by appropriately choosing a parameter q, such that, despite the statistical control signal, a specific set of operating points is selected at which equivalent measurements can be performed.

[0016] The values ​​r and q can be any numbers, as long as they can be processed by a standard computer system. For example, r and q can be real numbers rounded to a specific decimal place. The word length of r and q in bits can also be arbitrary.

[0017] The value of q can be π. This causes the value of the control signal to fluctuate statistically between the value of the primary signal x and the value shifted by π (x – π). Thus, depending on the value of x, an operating point statistically jumps by half a period of a cosine or sinusoidal function whose phase is modulated by the phase modulator. The absence of a mean value ensures that no control signals with a low frequency component are fed to the phase modulator, thereby improving the determination of small rotation rates in a fiber-optic gyroscope. Furthermore, for a value of q equal to π, it is ensured that operating points with positive and negative slopes are selected in a statistically distributed manner.

[0018] The value of q can also be 2π. This causes the value of the control signal to fluctuate statistically between the value of the primary signal x and the value shifted by 2π (x – 2π). Thus, depending on the value of x, an operating point statistically jumps by one complete period of a cosine or sinusoidal function whose phase is modulated by the phase modulator. The absence of a mean value ensures that no control signals with a low frequency component are fed to the phase modulator, thereby improving the determination of small rotation rates in a fiber-optic gyroscope. Furthermore, for a value of q equal to 2π, it is ensured that operating points with the same slope are selected in a statistically distributed manner.

[0019] The primary signal can be a superposition of a square wave and a ramp signal. This ensures that the jumps between different values ​​of the square wave generate large phase shifts through the phase modulator, which could be used, for example, to shift operating points. The ramp signal is superimposed on the square wave and guarantees a constant phase shift, thus maintaining an operating point.

[0020] The control signal can have a word width that is 1 bit wider than the word width of the primary signal. This allows the statistical, averaging-free control signal to be easily generated from the primary signal by adding a bit containing the random information. This enables the simplest possible data transmission and data structure.

[0021] The phase modulator can include a multifunctional integrated optical chip (MIOC) that receives the control signal. In such a known MIOC, the phase of passing light is modulated using electrodes. An electric field applied via the electrodes can influence the effective index, or the ability to conduct light. This allows for a particularly simple modulation of the phase of passing light.

[0022] A fiber optic gyroscope can incorporate the control system described above. Furthermore, the fiber optic gyroscope can include a light source for emitting light with a specific wavelength and a beam splitter for splitting the light from the light source into two incoming beams, which are directed into the phase modulator, and for superimposing the two outgoing beams exiting the phase modulator to form a detection beam. Additionally, the fiber optic gyroscope can include a coil for rotating the incoming beams received in the phase modulator in opposite directions before they are fed back into the phase modulator as two outgoing beams, and a detector for measuring the intensity of the detection beam, which is then transmitted to the control unit. The control unit uses the measured intensity to determine a rotation rate around a central axis of the coil and the statistical, mean-free control signal.

[0023] In the fiber optic gyroscope, a light source generates a light beam, i.e., a light signal, with a specific wavelength, for example, a laser beam with a predetermined wavelength. This beam is coupled into an optical fiber and fed to a beam splitter, which divides it into two beams that enter the phase modulator. As explained above, this can be a MIOC (multi-interface optical circuit). The beam splitter can also be integrated directly into the MIOC, for example, such that the MIOC has a Y-shaped optical path, i.e., two inputs / outputs on one side and one input / output on the other.

[0024] In the phase modulator, the phase of the two beams relative to each other is shifted by a value determined by the control signal before the beams are coupled into the coil. The coil is designed such that it consists of at least one optical fiber wound in a single plane. The beams are coupled into this optical fiber in such a way that the two beams circulate in opposite directions within the coil; that is, if one beam travels counterclockwise, the other travels clockwise.

[0025] If the coil is rotated around a central axis perpendicular to the plane of the windings while the light rays pass through it, the effective path of the light rays until they exit changes. The path of rays traveling against the direction of rotation is shortened, while the path of rays traveling with the direction of rotation is lengthened. Light rays that had the same phase upon entering the coil have a phase difference upon exiting, which is proportional to the rotation rate (Sagnac effect). This phase difference can be derived from the resulting interference pattern when the light rays are superimposed. In this case, the phase difference of the incoming rays generated by the phase modulator is therefore further altered by the phase difference resulting from the Sagnac effect.

[0026] To generate a readable interference pattern, the outgoing beams are coupled back into the phase modulator and the beam splitter, which superimposes the beams. In the phase modulator, an additional phase shift is modulated onto the beams, depending on the control signal. Due to the reciprocity of the setup, if the control signal remains constant, the phase shift between the oppositely circulating beams would disappear again after one revolution.

[0027] Phase modulation is generated to select specific operating points for measuring the interference signal. To enable the best possible determination of the rotation rate, the operating points are selected, for example, by being points of maximum slope on the interferometer characteristic curve, and by alternating between points of positive and negative slope. To avoid autocorrelations in the readout circuit of the fiber optic gyroscope, the switching between the operating points must not be periodic; there are four operating points.

[0028] Since the phase shifts when the light beams enter and exit the coil cancel each other out when the control signal remains constant, the control signal can be changed over time in such a way that the desired operating points are achieved.

[0029] The interference signal, modulated accordingly and consisting of the superimposed beams, is coupled into the detector via optical fibers, where its intensity is measured at the operating points determined by the phase modulation. From these intensity measurements, the control unit then determines the applied rotation rate and the control signal necessary for controlling the phase modulation. The phase modulator, detector, and control unit thus form a closed control loop.

[0030] The arrangement described above makes it possible to provide a fiber optic gyroscope which, due to control by a statistical, mean-free control signal, does not exhibit a lock-in effect at rotation rates near zero, which is caused by the frequency response of a phase modulator.

[0031] As explained above, the fiber optic gyroscope can use the phase modulation signal to modulate the phases of the two incoming and two outgoing beams in such a way that the intensity of the detection beam can be measured at predetermined operating points. This ensures optimal measurement conditions, thereby increasing the precision of the fiber optic gyroscope.

[0032] The operating points can be consecutive points of maximum slope of the detection beam intensity plotted against phase, and either 4 or 8 operating points can be used. As explained above, selecting at least four operating points (readout range of 2π) reduces the risk of autocorrelations in the evaluation circuit of the fiber optic gyroscope. Choosing 8 operating points, i.e., a readout range of 4π, allows for greater flexibility in setting parameters within the control loop of the fiber optic gyroscope. This makes it possible to fulfill all operating point requirements for high-precision operation of the fiber optic gyroscope, even at low rotation rates.

[0033] The fiber optic gyroscope can be a fiber optic Sagnac interferometer. This makes the advantages explained above accessible for measurements with a fiber optic Sagnac interferometer as well.

[0034] These and other advantages of the invention are explained below using examples and the accompanying figures. It shows:

[0035] Fig. 1A a schematic block diagram of a control system according to one embodiment;

[0036] Fig. 1B a schematic block diagram of a control system according to a further embodiment;

[0037] Fig. 2 a schematic block diagram of data transmission in a control system according to one embodiment;

[0038] Fig. 3 a schematic block diagram of a fiber optic gyroscope according to one embodiment;

[0039] Fig. 4 a schematic block diagram of a digital evaluation circuit according to the state of the art;

[0040] Fig. 5 a schematic block diagram of data transmission in a control system according to one embodiment;

[0041] Fig. 6 a schematic block diagram of data transmission in a control system according to a further embodiment; and

[0042] Fig. 7A to Fig. 7C Results of gyratory rate measurements using a prior art control system and control systems according to embodiments of Fig. 5 and Fig. 6.

[0043] Fig. 1A shows a control system 100 , which is suitable for controlling a fiber optic gyroscope. The control system 100 features a phase modulator 110 and a control unit 120 on.

[0044] The phase modulator 110 is suitable for determining the phase of a signal 115 , which it undergoes to change. With the phase modulator 110 For example, it is possible to shift the phase of a passing light beam by a specific amount. The phase modulator 110For example, it can incorporate an Integrated Optical Multifunction Chip (MIOC). In the MIOC, the phase of passing light is modulated using electrodes. An electric field applied via these electrodes can influence the effective index, or the ability to conduct light. This allows for a particularly simple way to modulate or shift the phase of passing light.

[0045] From the control unit 120 The phase modulator 110 a control signal 125 supplied, with which the phase modulation is controlled. The control unit 120 In addition to providing the control signal, it can 125 It can also take on other tasks. For example, the control unit can 120a computer processor, such as a CPU, which performs all or part of the processes in a system or device, such as a fiber optic gyroscope, in which the control system 100 is used, controlled.

[0046] The strength of the phase modulation is determined by the size of the control signal. 125 determined. For example, in MIOC, the control signal can be the voltage applied to the electrodes, and the phase shift can be directly proportional to the voltage value. The proportionality constant is the electro-optical gain factor of the modulation structure used. This can be determined for the modulation structures used. Scaling the control signal 125 with the inverse of the electro-optical gain factor, the magnitude of the resulting control signal is 125 that is equal to the phase shift, i.e., the value of the control signal. 125directly indicates the phase shift. The compensation of the electro-optical gain factor can be achieved, for example, in the control unit. 120 The digital-to-analog (DA) conversion is carried out.

[0047] The control signal 125 The signal used to control phase modulation is a statistical signal. This means that, unlike deterministic signals, its value is not uniquely determined but follows a specific probability distribution. In each processing cycle of the control unit, 120 The value of the control signal is thus determined 125 from a set of values, each assigned a probability of occurrence. The control signal therefore exhibits a statistical, i.e., unpredictable, fluctuation.

[0048] The statistics of the control signal 125 is characterized by the fact that the mean value of the control signal 125is equal to zero. The control signal 125 It is therefore mean-free. From these properties of the control signal... 125 This results in the frequency spectrum of the control signal 125 Frequencies near zero are strongly suppressed. This makes it possible to use the control signal 125 To operate phase modulators, such as MIOCs, whose response at low frequencies differs from their response at high frequencies. For example, lower electro-optical gain in the MIOC at low frequencies can be made imperceptible by using a statistical, averaging-free control signal. 125 The signal used is one that exhibits no or only strongly suppressed frequency components in the frequency range amplified less by the MIOC. This allows a phase modulator with different response characteristics at low and high frequencies to be operated stably and reliably.

[0049] The control signal125 This can be generated from a primary signal that is itself deterministic. The primary signal therefore takes a value from the control unit. 120 assigned value based on which the control signal is then sent 125 is generated. This results in the control signal being free of mean values. 125 another property of the control signal 125 specified, namely the basis on the primary signal. This allows the phase modulator 110 To control effectively, the primary signal is used as the original signal to set the phase modulation to a specific value, which is then modified to create a statistical, mean-free control signal. 125 results.

[0050] The primary signal can, for example, be a superposition of a square wave signal with at least four temporarily fixed signal values ​​and a ramp signal. The temporarily fixed signal values ​​of the square wave signal are then used in the phase modulator. 110 The first method involves creating a sudden, discontinuous phase shift by jumping between different signal values. In contrast, the ramp signal is used to generate continuously increasing phase shifts or to apply an external factor to the signal. 115 to compensate for the resulting phase shift.

[0051] This example shows Fig. 1B a schematic block diagram of the control system 100 , in which the control unit 120 a random number generator 130 The random number generator 130It generates random numbers r in the range from 0 to q, where q can be any real number. The number of digits for the numbers r and q is limited only by the processing power of the control unit. 120 or limited by the word length of the numbers r and q. Every random number r can be generated by the random number generator with a probability of 1 / q (or q for q < 1).

[0052] The control unit 120 The generated primary signal can have the value x and vary over time. Depending on the time-varying magnitude of the statistical random number r, the deterministic primary signal x, and the time-constant number q, the control unit... 120 the statistical quantity x r formed r x takes the value 0 for a value of (x + r) less than q, and the value q for a value of (x + r) greater than or equal to q. r is the statistical rounding of the primary signal x.

[0053] The control signal 125 is derived from the primary signal x and the value x r generated and has the size (x – x r ). Due to the probability distribution of the random number r and the rounding quantity q, it can easily be verified that the control signal 125 is mean-free. The magnitude of the control signal 125 The result is therefore, depending on the magnitude of the primary signal x and the random number r, either equal to the primary signal x or reduced by the amount q relative to the primary signal x. The larger the (deterministic) primary signal x, the more probable the control signal. 125 reduced by the value q.

[0054] This strongly suppresses low frequencies contained in the primary signal x, since the statistical switching between the values ​​x and (x – q) leads to an increase in the frequency of the primary signal x.

[0055] In a fiber optic gyroscope, phase modulation is used to perform the measurement necessary for determining the rotation rate at specific operating points of an interferometer characteristic curve, which indicates the interference of two light beams superimposed in the fiber optic gyroscope.

[0056] There are certain degrees of freedom regarding the selection of the sequence of operating points, which must be used effectively. Controlling these operating points via the phase modulator pursues the following objectives: 1. Without modulation, the apexes of the cosine inferometer characteristic curve, which have a slope of zero, would be targeted. Thus, the sensitivity of the fiber optic gyroscope would be zero, and no directional information would be available. To avoid these disadvantages, points with the steepest slope are targeted. 2. If only points with the same sign were driven, an applied rotation rate would result in a DC voltage signal that would be suppressed by subsequent amplification. Therefore, operating points with alternating signs are driven. This creates a readout signal that lies within the passband of the subsequent amplifier stages. 3. If the operating points were controlled in such a way that positive and negative slopes alternate periodically, a correlation between the control signal would occur. 125 and other signals necessary for the operation of the fiber optic gyroscope, which would lead to insensitivity zones at low rotation rates. Therefore, the sequence of signs of the slopes of the operating points (modulation) must be such that this correlation becomes zero. 4. The modulation must be designed so that the electro-optical gain factor can be compensated as described above for any input rotation rates of the sensor.

[0057] If q = π is selected, the phase changes statistically by the value π, resulting in a switch to an operating point with the opposite slope and thus changing the demodulator reference. For q = 2π, the phase changes statistically by the value 2π, while the sign of the slope remains unchanged.

[0058] Fig. Figure 2 shows a circuit-based implementation of generating the mean-free value (expected value E(x – x)). r ) = 0) of the control signal 125 The primary signal can be time-dependent and from the range x(t) ∊ [0, q). For example, it can have a word length of 12 bits, where the weighting of the most significant bit (MSB) is q / 2. This signal x(t) is then used in the adder. 140 the random number generator 130A uniformly distributed random number r, whose word width can also be 12 bits, is added to the generated signal. The condition x + r ≥ q is signaled by the occurrence of a carry. The sum itself need not be used. If the 12 bits of the signal x are supplemented with this carry as a new MSB and the resulting 13-bit number is interpreted as a two's complement, then the control signal obtained in this way is... 125 mean-free, since the new MSB (carryover) has the weighting -q.

[0059] The primary signal can also have a different word length. In that case, the primary signal is again modified by the carry amount to become the control signal (MSB). 125 added, which therefore has one more bit than the primary signal.

[0060] The control signal 125 with value (x – x r ) and E(x – x r ) = 0 can now be used with a DA converter 150 to be supplied, which in turn is the phase modulator 110It controls. Different values ​​of q are possible here, e.g.: 1. q = π and consequently a control range of the phase modulator 110 of 2π (hereinafter referred to as 2π modulation) 2. q = 2π and consequently a control range of the phase modulator 110 of 4π (hereinafter referred to as 4π modulation).

[0061] According to another embodiment, the control system described above 100 used in a fiber optic gyroscope for determining rotation rates, e.g. in a Sagnac interferometer. Fig. Figure 3 shows such a fiber optic gyroscope. 200 .

[0062] The structure of the fiber optic gyroscope 200 in Fig. 3 corresponds to the commonly used arrangement. A light source 201 emits light of wavelength λ and frequency ω = 2πc / λ, where c is the speed of light.

[0063] The light waves pass through a coupler. 202 and are then divided into a beam splitter 203 The beams are split into two partial beams. Both partial beams pass through a phase modulator. 210 , which imposes an additional phase modulation on them. This results in a phase shift between the two beams –φ(t) = –c1·u φ (t). Here u φ a control voltage of the phase modulator 210 and c1 is its electro-optical gain factor. The negative sign for the resulting phase difference was chosen arbitrarily.

[0064] Subsequently, both beams pass in opposite directions through a coil rotating at an angular velocity Ω against inertial space. 204 Fiber wound with radius R and total length L0. Due to the Sagnac effect, a further phase shift φ occurs. s= Ω·S, with S = 4πRL0 / (λc) acting between the two beams. The travel time of the light through the fiber coil is T0. After both beams have passed through the coil 204 Having passed through the phase modulator, the phase shift between them is Ω(t)·S – φ(t – T0). Both beams now pass through the phase modulator again. 210 , but this time with the function reversed, so that the phase φ(t) with a positive sign is added as a further component. The two from the phase modulator 210 The outgoing beams are contained in the beam splitter. 203 Therefore, the total phase shift Ω·S + φ(t) – φ(t – T0) is brought to interference.

[0065] After the merging, the light wave travels back to the coupler as a detection beam. 202 , where part of the detection beam goes to the detector 205 is guided. There, an output voltage u is generated that depends on the phase shift of the interfering light rays. det= c0cos(Ω·S + φ(t) – φ(t – T0)). The constant c0 depends on the average light power at the receiver, its sensitivity, and the gain of subsequent stages.

[0066] The remaining circuit section in Fig. 3 serves as a control unit 220 and has the purpose of feeding suitable signals into the phase modulator 210 the Sagnac interferometer 200 to bring into a state that allows evaluation of the detector signal u det permitted for the purpose of determining the rotation rate Ω.

[0067] The detector 205 generated signal u det a first amplifier stage 221 The signal is fed in with adjustable gain a0. This raises the signal to a defined level a0u. det brought and then through an AD converter 222 digitized. The resulting signal x AD a digital evaluation circuit 223supplied, which in turn generates a signal y DA This output signal, corresponding to the primary signal, is generated by a digital-to-analog converter (DAC). 224 translated into an analog voltage, and after multiplication by an adjustable gain factor a1 at a second gain stage 226 the phase modulator 210 supplied. For gain adjustment, a multiplying DA converter is expediently provided, whereby its reference voltage is used to influence the gain.

[0068] Typically, the DA converter 224 and the second amplifier stage 226 The electro-optical amplification factor c1 compensates.

[0069] The digital evaluation circuit 223 and the DA and AD converters 222 , 224 working with the cycle time T0, the transit time of the light through the coil 204Therefore, a closed signal path exists. The digital evaluation circuit 223 delivers output sizes y at specific, selectable times. Ω for the rotation rate, y a0 for the gain factor a0 of the input branch and y a1 for the gain factor a1 of the output branch. All these values ​​are averages that are sent to a processor for further processing. 227 will be provided. In addition, the digital evaluation circuit will be made available. 223 from the processor 227 or a timer after each reading of the averaged output values ​​with a "clear" command, which serves to reset the internal average value images.

[0070] The processor 227 calculated from the pre-averaged values ​​y Ω , y a0 and y a1After possible further filtering, the measured quantity Ω and the digital signals necessary for setting the gain factors a0 and a1, which are passed through a first auxiliary DA converter. 228 and a second auxiliary D / A converter 229 the corresponding first and second amplifier stages 222 , 224 influence.

[0071] A digital evaluation circuit 223 according to the state of the art, in Fig. 4 outlined.

[0072] Since a relationship exists between the digital data words y via a1 and c1 DA The optical phase φ can be established by a suitable choice of a1 such that a1c1 = 1, so that the individual bits in the data word y corresponding to the primary signal are DA Phase shifts φ at the modulator of size π·2 k correspond. To simplify further explanations, these values ​​will be called γ. k = π·2 kdirectly assigned to the place values ​​of the bits of the digital data word. This agreement is intended to apply except for y. DA for all digital data words of the digital evaluation circuit 223 apply, i.e., also to the data words s1, i = 1, ..., 8, s ' 3 , s ' 5 , y α0 , y α1 and y Ω out of Fig. 4. This means that, contrary to convention, the numerical value of a data word s is represented by the bits α k , k = l, ..., m according to calculated. Here, α1 is the "least significant bit" (LSB) and α m The MSB of the data word. For the data word y DA with the bits α ' k , k = l' ... m' is then

[0073] Because φ = a1c1y DA For a1c1 = 1, the phase shift in the phase modulator is 210 φ = y DA Therefore, in this case, the following applies:

[0074] As shown, m' = 0.

[0075] The AD converter 222supplied input signal x AD The internal signal s1 is fed to an input of the adder ADD1. This depends on a demodulation signal. d ' 2 , that can take the values ​​0 or 1, another evaluation with 1 – 2d ' 2 , so performed with +1 or -1. The demodulation signal d ' 2 (i) is the modulation signal d2(i) supplied by a random number generator M, delayed by n clock cycles by a delay V2, i.e. d ' 2 (i) = d 2 (i – n). The value n can be preset within predefined limits and serves to adapt the propagation time to the external signal path. The signals d2 and d2, respectively, are used for this purpose. d ' 2 Each can assume two states (0 or 1). For d ' 2 = 0 In stage ADD1, an addition takes place for d ' 2 = 1 A subtraction of the quantity s1. The other input of the adder is connected to a register pair RP1, which stores two predefined values ​​+d and -d. The test quantity ±d is, as will be shown later, fed into the main control loop as an additional signal, with the aim of "measuring" its loop gain and adjusting it to a defined setpoint using an auxiliary control loop that influences a controllable amplifier. The test signal ±d, superimposed on the useful signal, should be chosen to be sufficiently small to avoid overloading the external gyro path. With correctly set gain, as will be shown, this test signal is exactly compensated, so that the measurement accuracy of the sensor remains unaffected. A "select" input s is provided for selecting the desired value, which is controlled by a signal. d ' 1 is controlled. The selected value, effective at the adder input, is (2d ' 1 – 1)·d . This means s 2 (i) = 2d ' 1 (i) – 1)·d – (2d ' 2 (i) – 1)·s 1 (i)

[0076] The signal d ' 1 arises, analogous to d ' 2 , by n-stage delay using V1 from the signal d1. The signal d1 is generated by a random number generator D independent of M. The sum s2 generated by ADD1 is fed to the inputs of two averaging screens with ADD5 and ADD6, described below, and to the input of an adder ADD2. Its sum output s3 is fed to a register chain REG1 and fed back to the other adder input as signal d1 delayed by n clock cycles. s3(i) = s3(i – n) + s2(i)

[0077] In addition, s3 feeds the average value image with ADD7, which is explained further below, as well as the adder ADD3. The signal d2, described above and supplied by the random number generator M, is fed into the other input of ADD3 with a value of π. The selectable output of a register pair RP2, containing the stored values ​​π / 2 + d and π / 2 – d, is connected to the smaller-valued digits (π / 2, π / 4, ...) of the same input. Selection is performed using the signal d1 described above, which is generated by the random number generator D. This ensures that s4(i) = s3(i) + π / 2 + d 2π + (2d1 – 1)·d

[0078] From the sum signal s4 of the adder ADD3, all bits with a value of 2π and greater are now separated at position "tr". This process corresponds to a modulo-2π operation.

[0079] The remaining bits are fed to the input of the phase integrator consisting of ADD4 and REG2. The sum output s5 of ADD4 also contains only bits with values ​​less than 2π. It is delayed by one clock cycle by REG2 and fed back to the other adder input. The carry bit C resulting from the addition is fed as signal d3 to the delay chain V3. This ensures that

[0080] Simultaneously, the output of REG2 serves as the primary signal output signal y DA led to the outside, to the DA converter.

[0081] As mentioned above, signals s2 and s3 are processed using three averaging plots. These are externally resettable accumulators that sum the signal to be averaged over a predetermined time period of m clock cycles.

[0082] The average rotation rate value y Ω arises through the accumulation of s3 with ADD7 and REG5:

[0083] The control variable y a0 This results from an accumulation of s2 carried out with ADD5 and REG5, where an additional, by d ' 1 dependent weighting of s2 with +1 or -1 is performed:

[0084] Accordingly, y is created. a1 by d ' 3 dependent, weighted accumulation of s2 with ADD6 and REG4. d ' 3 is the signal d3 delayed by V3 by n clock cycles, which is formed from the carry bit C with the weight 2π of the adder ADD4 of the phase integrator:

[0085] Let us first assume that the factors a0 and a1 are set such that a0c0 = 1 and a1c1 = 1. Furthermore, due to the properties of the converters, n – 1 dead times must be taken into account. Then x AD (i + n) = cos(Ω·S + y DA (i + 1) – y DA (i))

[0086] As in Fig. 4 shown applies y DA (i) = s ' 5 (i) and y DA(i + 1) = s5(i). Furthermore, the following holds: s 4 (i) = s 5 (i) – s ' 5 (i) + k·2π

[0087] The k·2π-fold deviation results from the modulo-2π operation that occurs at "tr". The term k·2π can be omitted from the argument of the cosine function due to its periodicity. Therefore, s1(i + 1)x AD (i + n) = cos(Ω·S + s4(i))

[0088] First, let d = 0 in both register pairs RP1 and RP2. Then: s4(i) = s3(i) + π / 2 + d 2π ) and because cos(x + π / 2) = –sin(x) as well as sin(x) = –sin(x + π) and s1 = x AD : s1(i + n) = sin(Ω S + s3(i)) (2d2(i) – 1)

[0089] On the other hand, s 2 (i + n) = –s 1 (i + n)·(2d ' 2 (i + n) – 1) = s 1 (i + n)·(2d 2 (i) – 1)

[0090] From this it follows that s2(i + n) = –sin(Ω S + s3(i))

[0091] The digital evaluation circuit 223is a closed-loop control system that attempts to minimize the control deviation (ΩS + s3(i)). If this quantity, which is in the argument of the sine function, is small, the sine function can be approximately replaced by its argument, and the following holds true: s2(i + n) = –Ω S – s3(i) or, in z-transformed form: s2(z) = –z –n (Ω·S + s3(z))

[0092] The following stage, composed of ADD2 and REG1, with the transfer function This closes the control loop. Eliminating the quantity S2(z) from the last two equations yields the relationship... S3(z) = –z –n Ω·S

[0093] The signal s3 is therefore proportional to the rotation rate Ω. The average image consisting of ADD7 and REG5 generates the signal y from this. Ω .

[0094] The preceding explanation assumes that the condition a1c1 = 1 is satisfied. A special auxiliary control loop is intended to adjust a1 until this requirement is met. This utilizes the fact that the digitally performed modulo-2π operation generates an additional error signal if the phase in the interferometer does not jump by exactly the value 2π corresponding to the modulo operation. The phase effective at the phase detector is φ d (i + 1) = Ω S + a1c1(s5(i) – s5(i – 1))

[0095] If the product a1c1 deviates from the ideal value of 1, a phase error is added to the "ideal" detector phase. φ e (i + 1) = (a1c1 – 1)(s5(i) – s5(i – 1)) In addition, after demodulation, this phase error appears as an additional rotation rate signal. This error signal is therefore the scale factor deviation modulated with s5(i) – s5(i – 1). However, it holds that s5(i) – s5(i – 1) = mod2π[s4(i)] – 2πd3(i)

[0096] The right-hand side of this equation can be interpreted as a two's complement number with the sign bit d3. Thus, d3 is the sign of the signal s5(i) – s5(i – 1) modulating the scale factor deviation (a1c1 – 1). The error modulated in this way appears at node s2 after n clock cycles and can be determined by its sign, which is also delayed by n clock cycles. d ' 3 (i) The signal must be demodulated to derive a control variable for a1. This is done using the average value pattern constructed from ADD6 and REG4. Additional demodulation is performed via the ± control input of the adder. The averaged signal is then displayed at output y. a1 is therefore a measure of the deviation of the factor a1 from the target value (given by a1c1 = 1) and is used to adjust the factor to the target value.

[0097] For the stability of the main control loop, it is necessary that the loop gain has the correct value, determined by a0c0 = 1. To ensure this condition is always met, an auxiliary control loop is provided for setting a0. For Ω = 0s2, the signal delayed by n clock cycles is -s3. For Ω = 0 and a0c0 ≠ 1, the following applies: s2(i + n) = –a0c0s3(i)

[0098] To automatically find a measure for the deviation of factor a0 from the ideal value, a small test quantity +d and -d is stored in the register pair RP2 in addition to the value π / 2. This injects an additional test signal 2d1(i) – 1·d, whose sign is controlled by the random number generator D, into the adder ADD3 to s3. If one is only interested in the effect of the test signal alone, then the following applies: s2(i + n) = –a0c0·(2d1(i) – 1)·d

[0099] If the same test values ​​+d and –d are now stored in the register pair RP1, then the test signal (2d1(i + n) – 1)·d is added to s2(i + n) and the following applies: s2(i + n) = (1 – a0c0)(2d1(i) – 1) d

[0100] Therefore, a component of the test signal weighted with (1 – a0c0) is present at the node. This component is filtered out using the averaging images ADD5 and REG3, whose input signal s2 is additionally weighted with the sign of the test signal. This results in the averaged signal y. a0 a measure of the deviation of the product a0c0 from 1 and can be used to adjust a0 to its target value.

[0101] With the control unit described above 220 integrated digital evaluation circuit 223 It is therefore possible to determine both the rotation rate Ω and the parameters necessary for its calculation. Furthermore, the evaluation circuit provides 223 The primary signal is the signal yAD This signal is directly incorporated into the phase modulation. This primary signal is neither static nor averaging-free.

[0102] A possible modification of the digital evaluation circuit 223 according to the example of the circuit-technical implementation from Fig. 2 is in Fig. 5 shown. Fig. Figure 5 shows the elements D, ADD3, ADD4 and REG2 from Fig. 4, which output the signal y DA are needed.

[0103] Instead of the signal y DA but directly to the DA converter 224 To pass on the signal y DA using a random number generator 230 and an adder 240 as above in Fig. 2 described, modified. The signal y DA This corresponds to the primary signal x.

[0104] The in Fig. The circuit shown in section 5 is used for a control range of the phase modulator. 210 used for 2π, i.e. for q = π.

[0105] Fig. Figure 6 shows another possible implementation of generating a statistical, mean-free control signal based on the digital evaluation circuit known from the prior art. 223 Here too, the signal y DA (Primary signal x) by the random number generator 230 and the adder according to the above at Fig. The two described methods were modified to produce a statistical, mean-free control signal.

[0106] Unlike the circuit from Fig. 5 is in the circuit from Fig. 6. The operating range of the phase modulator is 4π, i.e., q = 2π. This provides additional degrees of freedom in selecting the operating points and control loop parameters compared to the setting q = π, thus significantly simplifying the circuit design. For example, for q = 2π, it is not necessary to use the random number generator. 230 and adder 240generated signal, as in the case q = π from Fig. 5, via a gate consisting of several XOR elements 560 into the digital evaluation unit 223 to be attributed to.

[0107] In the Fig. 7A to Fig. Section 7C schematically shows, using diagrams, how the measurement accuracy of a fiber optic gyroscope at zero rotation rates is improved by using a control system that employs a statistical, mean-free control signal for phase modulation.

[0108] Fig. Figure 7A shows the rotation rate in ° / h versus time for a fiber optic gyroscope with a state-of-the-art control system at a rotation rate of 0° / h. In a perfectly functioning system, one would expect only statistically distributed noise, meaning the rotation rate values ​​would be Gaussian distributed. However, as can be clearly seen, the measurement curve shows constrictions at the value of 0° / h, indicating a lock-in effect that systematically distorts the measurement. The measurement results are no longer purely Gaussian distributed. This is also evident in the lower diagram of the Fig. Figure 7A clearly shows the frequency distribution for different measured values. It is evident that this is not pure, Gaussian noise, as there is a significant peak in the distribution at 0° / h.

[0109] Fig. 7B and Fig. 7C shows the same diagrams for the circuits from Fig. 5, or Fig.6, i.e., for a statistical, mean-free control signal generated as described above with q = π, or q = 2π. It is clearly evident that the measurement is now only influenced by statistical noise; that is, the frequency distribution of the measured values ​​follows a Gaussian curve centered at 0° / h, and the measured values ​​no longer show any constrictions at 0° / h. Similar improvements are also observed when comparing measurements performed at low rotation rates, such as in gyrocompassing, where components of the Earth's rotation rate of 15° / h must be measured as accurately as possible.

[0110] It is therefore possible to reduce the lock-in effect at 0° / h and thereby increase the reliability of the measurement results at rotation rates near 0° / h by using a control system for a fiber optic gyroscope that generates a control signal for phase modulation which is statistical and free of mean values.

Claims

[1] Tax system ( 100 ) for a fiber optic gyroscope with: a phase modulator ( 110 ) to modulate the phase of a light signal ( 115 ); a control unit ( 120 ) to generate a control signal ( 125 ), by whose size the phase is modulated and which is the phase modulator ( 110 ) is supplied; characterized by that the control signal ( 125 ) changes statistically and is mean-free. [2] Tax system ( 100 ) according to claim 1, wherein the control unit ( 120 ) the mean-free control signal ( 125 ) generated from a deterministic primary signal. [3] Tax system ( 100 ) according to any of the preceding claims, wherein the control unit ( 120 ) a random number generator ( 130 ) exhibits a random number r between 0 and q with a probability of 1 / q; the control unit ( 120 ) generates a primary signal of size x; the control signal ( 125 ) the size (x – x r has; and It is true that x r = 0 for x + r < q and x r = q for x + r ≥ q. [4] Tax system ( 100 ) according to claim 3, wherein q = π. [5] Tax system ( 100 ) according to claim 3, wherein q = 2π. [6] Tax system ( 100 ) according to any one of claims 2 to 5, wherein the primary signal is a superposition of a square wave signal and a ramp signal. [7] Tax system ( 100 ) according to one of claims 2 to 6, wherein the control signal ( 125 ) has a word width that is 1 bit wider than the word width of the primary signal. [8] Tax system ( 100 ) according to any of the preceding claims, wherein the phase modulator ( 110) has an integrated optical multi-function chip that transmits the control signal ( 125 ) receives. [9] Fiber optic gyr ( 200 ) with a tax system according to one of the preceding claims and with: a light source ( 201 ) to emit light with a specific wavelength; a beam splitter ( 203 ) to divide the light from the light source ( 201 ) into two incoming beams that enter the phase modulator ( 210 ) are directed, and used to superimpose two signals from the phase modulator ( 210 ) emerging outgoing rays to a detection beam; a coil ( 204 ) for the counter-rotating of the phase modulator ( 210 ) received incoming rays before they reach the phase modulator as two outgoing rays ( 210 ) are reintroduced; a detector ( 205) to measure the intensity of the detection beam, which is sent to the control unit ( 220 ) is passed on; whereby the control unit ( 220 ) from the measured intensity a rotation rate around a central axis of the coil ( 204 ) and determines the statistical, mean-free control signal. [10] Fiber optic gyr ( 200 ) according to claim 9, wherein the phase modulator ( 210 ) is suitable to modulate the phases of the two incoming beams and the phases of the two outgoing beams in such a way that the intensity of the detection beam can be measured at predetermined operating points. [11] Fiber optic gyr ( 200 ) according to claim 10, wherein the operating points are successive points of maximum slope of the intensity of the detection beam plotted against the phase and 4 or 8 operating points are used. [12] Fiber optic gyr ( 200) according to one of claims 9 to 11, wherein the fiber optic gyroscope ( 200 ) is a fiber-optic Sagnac interferometer.

Images