A method for online calibration of a hemispherical resonator gyro based on voltage fluctuation

Abstract

The application relates to an online calibration method for a hemispherical resonator gyro based on voltage fluctuation, and belongs to the technical field of inertia. The method solves the problems of poor real-time calibration performance of a traditional hemispherical resonator gyro virtual precession speed online calibration method under full temperature, and the problems of dependence on experimental data of previous calibration, trouble, labor and poor environmental adaptability. The application takes precession error convergence as a condition for triggering the opening of precession error calibration, adjusts the virtual precession control voltage V according to the precession error obtained by the i-th round of identification and beta 20 Adjust the virtual precession control voltage V w0 Make the hemispherical resonator gyro precess in the form of a constant value superimposed on a specific frequency reference signal under the virtual precession control force corresponding to the adjusted virtual precession control voltage, so that the precession speed of the standing wave remains constant, and the online calibration of the hemispherical resonator gyro is completed. The application is applied to the online calibration of the hemispherical resonator gyro virtual precession.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of inertial technology. BACKGROUND

[0002] The hemispherical resonator gyroscope is a Coriolis vibration gyroscope that detects angle or angular velocity by using the Coriolis effect. The working mode of the hemispherical resonator gyroscope includes a full-angle mode and a force balance mode. The full-angle mode refers to the fact that the azimuth angle of the resonator is always in a free precession state, and the force balance mode is to fix the azimuth angle at 0°. As a kind of solid-state vibration gyroscope, the hemispherical resonator gyroscope exhibits high reliability, long service life and strong anti-radiation stable physical properties, and is widely used in military fields such as aviation, aerospace and missile systems. The performance of the hemispherical resonator gyroscope is affected by the gyroscope assembly level, resonator structure error and circuit error. In order to eliminate the influence of the above factors on the drift of the gyroscope during the operation of the gyroscope, a virtual precession is usually used to realize online identification and compensation of error factors.

[0003] However, the virtual precession of the full-angle mode hemispherical resonator gyroscope adopts an open-loop mode, that is, the azimuth angle of the resonator is always in a free precession state, and the virtual precession control voltage is constant, so that the virtual precession speed changes regularly with the change of frequency and high voltage. The virtual precession speed is the change of the standing wave precession speed under the premise of no external angular velocity input by using the virtual precession control voltage, and at this time the precession speed of the standing wave is called the virtual precession speed.

[0004] When the gyroscope changes in the environment temperature, the resonant frequency and the direct current high voltage change rapidly, and then the virtual precession needs a long time to make the standing wave precess at a constant speed under the action of the virtual precession control voltage and reach stability. On the one hand, the calibration of the virtual precession in the prior art is to collect a large amount of experimental data to generate a lookup table, and to dynamically call the virtual precession control voltage according to the resonant frequency. The whole process is time-consuming, laborious and poor in environmental adaptability. This method is very dependent on the accuracy of the model, and thus the compensation effect is very limited.

[0005] On the other hand, the resonant frequency and the virtual precession speed are calibrated and fitted by using a temperature box in the prior art, which needs to rely on external experimental equipment and a large amount of calibration experiments in the early stage.

[0006] Therefore, in order to improve the real-time calibration performance of the gyroscope at full temperature, overcome the problems of the traditional hemispherical resonator gyroscope virtual precession speed online calibration method, such as time-consuming, laborious and poor environmental adaptability, it is very meaningful to propose a hemispherical resonator gyroscope virtual precession online calibration method based on voltage fluctuation. SUMMARY

[0007] The application aims to solve the problems of poor real-time calibration performance at full temperature, dependence on pre-marking experimental data, and poor environmental adaptability of the conventional hemispherical resonator gyro virtual precession speed online calibration method, and provides a hemispherical resonator gyro virtual precession online calibration method based on voltage fluctuation.

[0008] A hemispherical resonator gyro virtual precession online calibration method based on voltage fluctuation, which comprises the following steps:

[0009] Step 1, power on the hemispherical resonator gyro to start, so that it is at the initially given virtual precession control voltage V w0 The corresponding virtual precession control force F w0 The lower precession;

[0010] Step 2, use the same frequency cosine reference signal output by the phase-locked loop to demodulate the vibration signal detected in real time by the X and Y channels of the hemispherical resonator gyro, to obtain the X channel cosine signal C x And the Y channel cosine signal C y ;

[0011] Step 3, angle calculation is performed on C x And C y , to obtain the standing wave azimuth angle θ; then differential operation is performed on the standing wave azimuth angle θ to obtain the standing wave precession speed

[0012] Step 4, i-th round of identification, the initial value of i is 1, the initial value of the precession error is set as β 20 , according to the standing wave precession speed And the specific frequency reference signal, the precession error is identified online by using the precession error identification algorithm;

[0013] Step 5, judge whether the precession error obtained in the i-th round of identification converges, if yes, jump to step 6, otherwise, i=i+1, and jump to step 4;

[0014] Step 6, record the initially given virtual precession control voltage V w0 And the precession error obtained in the i-th round of identification, and prepare to start the precession error calibration;

[0015] Step 7, judge whether the power-on time of the converged hemispherical resonator gyro reaches the time threshold, if yes, jump to step 8, otherwise, jump to step 6;

[0016] Step 8, start the precession error calibration:

[0017] According to the precession error obtained in the i-th round of identification and β 20 Adjust the initially given virtual precession control voltage V w0the adjusted virtual precession control voltage V w0 corresponding to the virtual precession control force, the standing wave precesses in the form of a constant superimposed with a specific frequency reference signal, and the precession speed of the standing wave is kept constant, and the online calibration of the virtual precession of the hemispherical resonator gyroscope is completed.

[0018] Preferably, the time threshold is 10 minutes.

[0019] Preferably, the condition for determining whether the precession error obtained in the i-th round of identification converges in step 5 is:

[0020] When the difference between the precession errors obtained in the adjacent two rounds of identification is less than a set convergence threshold, it is determined that the precession error obtained in the i-th round of identification converges, otherwise, the precession error obtained in the i-th round of identification does not converge.

[0021] Preferably, the implementation of the X-channel cosine signal C x and the Y-channel cosine signal C y in step 2 is:

[0022]

[0023] wherein,

[0024] x = [acos2θcosωt-qsin2θsinωt];

[0025] y = [asin2θcosωt+qcos2θsinωt];

[0026]

[0027]

[0028] LPF(·) is a low-pass filter, V s and V c are the same-frequency sine and cosine reference signals respectively, x and y are the vibration signals detected by the X-channel and the Y-channel respectively, a and q represent the amplitude of the main standing wave and the amplitude of the orthogonal wave respectively, is the initial phase of the reference signal, ω is the resonant frequency of the resonator, and t is time.

[0029] Preferably, in step 3, the implementation of obtaining the azimuth angle θ of the standing wave is:

[0030] Preferably, the specific frequency reference signal is a given frequency sine signal.

[0031] Preferably, in step 4, the implementation of using the precession error identification algorithm to identify the precession error online includes:

[0032] Step 41, establishing the standing wave precession velocity and the initial given virtual precession control voltage V w0 the correlation model between them:

[0033]

[0034] wherein k is the precession factor of the half-sphere harmonic oscillator, Ω is the input angular rate, K d is the detection gain, K c is the control gain, a0 is the target amplitude, ω is the resonance frequency of the oscillator, V con-0 is the constant component, V sin is the amplitude of the sinusoidal component, λ is the frequency of the sinusoidal component, and t is the time;

[0035] Step 42, writing the correlation model in the form of discrete product:

[0036]

[0037] wherein:

[0038] α(i) = [1 sinλt(i)]

[0039]

[0040] is the expected output vector at the i-th sampling time, and is also the standing wave precession velocity at the i-th sampling time, i is the sampling time sequence number, i = 1, 2, 3, …, α(i) is the observation vector at the i-th sampling time, t(i) is the time corresponding to the i-th sampling time, β(i) is the parameter vector at the i-th sampling time, β1(i) is the first row and first column element in β(i), β2(i) is the second row and first column element in β(i), T is the matrix transpose, Ω(i) is the discrete quantity of Ω, K d (i) is the discrete quantity of K d , K c (i) is the discrete quantity of K c , V con-0 (i) is the discrete quantity of V con-0 , V sin (i) is the discrete quantity of V sin ;

[0041] Step 43, updating the gain vector H(i) and the covariance matrix P(i) of the precession error identification algorithm corresponding to the i-th sampling time according to α(i) in step 42:

[0042] Step 44, updating the gain vector H(i) and the covariance matrix P(i) of the precession error identification algorithm corresponding to the i-th sampling time according to H(i) in step 43, α(i) in step 42, and The parameter estimation β(i) at the i th sampling moment is obtained, and β2(i) in β(i) is taken as the precession error of the i th round of identification;

[0043]

[0044] Wherein, β(0)=[0 β 20 ] T .

[0045] Preferably, in step 43,

[0046]

[0047]

[0048] Wherein, P(i-1) is the covariance matrix corresponding to the i-1 th sampling moment; P(0)=I, I is a unit matrix, and η is a forgetting factor, which is selected in the range of (0, 1].

[0049] Preferably, in step 8, the expression of the adjusted virtual precession control voltage V w0 is as follows:

[0050]

[0051] Wherein, V con-0 is a constant component, β2(i) is the precession error of the i th round of identification, V sin is the amplitude of the sine component, λ is the frequency of the sine component, and t is time.

[0052] Preferably, in step 41, the value range of λ is 0Hz to 10Hz.

[0053] Advantages of the present application:

[0054] The present application superimposes a reference signal of a specific frequency in the virtual precession control signal of a gyroscope, and then uses the same-frequency cosine and sine reference signals output by a phase-locked loop and a recursive algorithm to realize online characterization of the precession error factor in the output angular rate of the gyroscope, and finally dynamically adjusts the virtual precession control voltage according to the identification result, thereby solving the problem of difficult calibration of the virtual precession of the gyroscope under different working conditions, significantly improving the stability and reliability of the gyroscope, and the present application can be applied to the fast start and online calibration process of the virtual precession of a hemispherical resonator gyroscope, saving time and effort, and being applicable to a full-temperature environment.

[0055] Compared with the method of calibrating and fitting compensation of resonant frequency and virtual precession speed by using a temperature box in the prior art, the present application superimposes a sinusoidal signal in the virtual precession control voltage of the gyro, then extracts the precession error from the sinusoidal signal of the virtual precession, and realizes online calibration of the virtual precession speed. The method does not need to rely on external experimental equipment, and does not need to perform a large number of calibration experiments in advance, and therefore can be widely applied to hemispherical resonator gyroscopes. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 is a connection relationship diagram of a control circuit with virtual precession control connected with a hemispherical resonator gyroscope;

[0057] Figure 2 is a principle schematic diagram of a hemispherical resonator gyroscope virtual precession online calibration method based on voltage fluctuation according to the present application;

[0058] Figure 3 is a flowchart of a hemispherical resonator gyroscope virtual precession online calibration method based on voltage fluctuation according to the present application. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0060] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0061] Reference is made to Figure 1 is a connection relationship diagram of a control circuit with virtual precession control connected with a hemispherical resonator gyroscope in the prior art; the control circuit includes a resonant frequency detection circuit, a virtual precession control circuit, a voltage fluctuation detection circuit, and a precession error extraction circuit. Figure 1 is a principle schematic diagram of a hemispherical resonator gyroscope virtual precession online calibration method based on voltage fluctuation according to the present application; w0The application solves the problem that the open-loop precession speed of the hemispherical resonator gyroscope is difficult to be characterized and calibrated in real time with temperature change, and overcomes the problem that the experimental data needs to be relied on the previous calibration, which is time-consuming, laborious and poor in environmental adaptability. The application superimposes a reference signal of a specific frequency in the virtual precession control signal of the gyroscope, and then uses the same frequency cosine reference signal and recursive algorithm to realize the online characterization of the precession error factor in the output angular rate of the gyroscope. Finally, the virtual precession control voltage is dynamically adjusted according to the identification result, so that the problem of difficult calibration of the virtual precession of the gyroscope in different working states is solved, and the stability and reliability of the gyroscope are significantly improved. The application can be applied to the rapid start and online calibration of the virtual precession of the hemispherical resonator gyroscope. The specific scheme is as follows:

[0062] Specific implementation Figures 1 to 3 In this embodiment, a kind of hemispherical resonator gyroscope virtual precession online calibration method based on voltage fluctuation is described, which includes the following steps:

[0063] Step 1, power on the hemispherical resonator gyroscope to start, so that it is in the initial given virtual precession control voltage V w0 The corresponding virtual precession control force F w0 Lower precession;

[0064] Step 2, the same frequency cosine reference signal output by the phase-locked loop is used to demodulate the vibration signal detected in real time by the X and Y channels of the hemispherical resonator gyroscope, to obtain the X channel cosine signal C x And Y channel cosine signal C y ;

[0065] Step 3, angle calculation is performed on C x And C y , to obtain the standing wave azimuth angle θ;Then, difference operation is performed on the standing wave azimuth angle θ to obtain the standing wave precession speed

[0066] Step 4, the i-th identification, the initial value of i is 1, and the initial value of the precession error is β 20 , according to the standing wave precession speed And the specific frequency reference signal, the precession error is identified online by using the precession error identification algorithm;

[0067] Step 5, judge whether the precession error obtained in the i-th identification is converged, the result is yes, jump to step 6, otherwise, i=i+1, jump to step 4;

[0068] Step 6, record the initial given virtual precession control voltage V w0 And the precession error obtained in the i-th identification, prepare to start precession error calibration;

[0069] Step 7, judging whether the power-on time of the converged hemispherical resonator gyroscope reaches a time threshold, and the result is yes, then jumping to step 8, otherwise, jumping to step 6;

[0070] Step 8, starting the precession error calibration:

[0071] According to the precession error obtained in the i-th round of identification and β 20 Adjusting the initially given virtual precession control voltage V w0 So that the hemispherical resonator gyroscope rotates at a constant value under the virtual precession control force corresponding to the adjusted virtual precession control voltage V w0 The standing wave rotates in the form of superimposing a specific frequency reference signal, and further makes the precession speed of the standing wave Constant, completing the online calibration of the virtual precession of the hemispherical resonator gyroscope.

[0072] In a specific application, when the precession error obtained in the i-th round of identification converges, it is proved that the precession error enters a stable state.

[0073] The specific frequency reference signal is a given frequency sine signal, wherein the frequency of the given frequency sine signal is usually within 10 Hz.

[0074] The present application superimposes a sine signal in the virtual precession control voltage, and further extracts the precession error from the sine signal of the virtual precession, and realizes the online calibration of the virtual precession speed. The present application method does not need to depend on external experimental equipment, and does not need a large amount of calibration experiment in advance, and therefore can be widely applied to the hemispherical resonator gyroscope.

[0075] Further, as an example in step 7, the value of the time threshold is 10 minutes. After the precession error converges, the gyroscope calibration is started after waiting for 10 minutes, so as to avoid the oscillation of the total energy E of the gyroscope and the quadrature Q of the gyroscope, and finally cause the precession speed of the standing wave to produce random jitter.

[0076] Since the virtual precession control loop needs to control the standing wave to rotate in the form of superimposing a specific frequency reference signal at a constant value, the initially given virtual precession control voltage V w0 Is expressed as:

[0077] V w0 = V con-0 + V sin sinλt;

[0078] V con-0 Is a constant component, which will be adjusted later to maintain the stability of the virtual precession speed, V sin Is the amplitude of the sine component, which remains constant, λ is the frequency of the sine component, and t is the time.

[0079] Further, the condition for judging whether the precession error obtained in the i-th round of identification converges is:

[0080] When the difference between the precession errors obtained in two adjacent rounds of identification is less than a set convergence threshold, it is determined that the precession error obtained in the i-th round of identification converges, otherwise, the precession error obtained in the i-th round of identification does not converge.

[0081] Further, the implementation of the X-channel cosine signal C x and the Y-channel cosine signal C y obtained in step 2 is as follows:

[0082]

[0083] wherein,

[0084] x = [acos2θcosωt-qsin2θsinωt];

[0085] y = [asin2θcosωt+qcos2θsinωt];

[0086]

[0087]

[0088] LPF(·) is a low-pass filter, V s and V c are the same-frequency sine and cosine reference signals respectively, x and y are the vibration signals detected by the X-channel and the Y-channel respectively, a and q represent the main standing wave amplitude and the orthogonal wave amplitude respectively, is the initial phase of the reference signal, ω is the resonant frequency of the resonator, and t is time.

[0089] Further, in step 3, the implementation of obtaining the standing wave azimuth angle θ is as follows:

[0090] Further, in step 4, the implementation of using the precession error identification algorithm to identify the precession error online includes:

[0091] Step 41, a correlation model between the standing wave precession velocity and the initially given virtual precession control voltage V w0 is established:

[0092]

[0093] wherein, k is the precession factor of the hemispherical resonator, Ω is the input angular rate, K d is the detection gain, K c is the control gain, a0 is the target amplitude, ω is the resonant frequency of the resonator, and Vcon-0 V is a constant component sin V is a constant component, λ is a frequency of the sinusoidal component, t is time, and λ is in the range of 0 Hz to 10 Hz in specific applications;

[0094] Step 42, write the correlation model into a discrete product form:

[0095]

[0096] Wherein:

[0097] α(i) = [1 sinλt(i)]

[0098]

[0099] is an expected output vector at the i-th sampling time, and is also a standing wave precession velocity at the i-th sampling time, i is a sampling time sequence number, i = 1, 2, 3, …, α(i) is an observation vector at the i-th sampling time, t(i) is a time corresponding to the i-th sampling time, β(i) is a parameter vector at the i-th sampling time, β1(i) is an element in the first row and the first column of β(i), β2(i) is an element in the second row and the first column of β(i), T is a matrix transpose, Ω(i) is a discrete quantity of Ω, K d (i) is a discrete quantity of K d (i) is a discrete quantity of K c (i) is a discrete quantity of K c (i) is a discrete quantity of V con-0 (i) is a discrete quantity of V con-0 (i) is a discrete quantity of V sin (i) is a discrete quantity of V sin (i) is a discrete quantity of V

[0100] Step 43, update the gain vector H(i) and the covariance matrix P(i) of the precession error identification algorithm corresponding to the i-th sampling time according to α(i) in step 42:

[0101]

[0102]

[0103] Wherein, P(i-1) is a covariance matrix corresponding to the i-1-th sampling time; P(0) = I, I is an identity matrix, η is a forgetting factor, and η is selected in the range of (0, 1]; when the forgetting factor is small, the algorithm converges fast, but the steady state is easily affected by noise; on the contrary, the larger the value of the forgetting factor, the slower the algorithm converges, but the steady state effect is not easily disturbed by noise;

[0104] Step 44, according to H(i) in step 43, α(i) in step 42, and The parameter estimator β(i) at the i-th sampling time is obtained, and β2(i) in β(i) is taken as the precession error of the i-th round of identification;

[0105]

[0106] wherein β(0) = [0 β 20 ] T .

[0107] Further, in step 8, the adjusted virtual precession control voltage V w0 is expressed as:

[0108]

[0109] wherein V con-0 is a constant component, β2(i) is the precession error of the i-th round of identification, V sin is the amplitude of the sinusoidal component, λ is the frequency of the sinusoidal component, and t is time. In specific applications, the value of λ ranges from 0 Hz to 10 Hz.

[0110] (I) In the early design stage, the design concept of the virtual precession control voltage V w0 comes from:

[0111] Referring to Figure 1 or Figure 2 Under ideal conditions, the vibration signals output by the two-channel preamplifier buffer circuit are:

[0112]

[0113] wherein x and y are the vibration signals detected by the X channel and the Y channel respectively, a and q represent the main standing wave amplitude and the orthogonal wave amplitude respectively, ω is the vibration frequency of the resonator, t is time, and θ is the azimuth angle of the standing wave.

[0114] The vibration signals detected by the X / Y channel are demodulated with the reference signal output by the phase-locked loop, and then the gyro control parameters E, Q, and L are obtained: E as the total energy of the gyro, Q as the orthogonal quantity of the gyro, and L as the phase-locked loop control quantity;

[0115] The reference signal output by the phase-locked loop is:

[0116] wherein V s and V c are the same-frequency sine and cosine reference signals, is the initial phase of the reference signal, ω is the resonant frequency of the resonator, and t is time.

[0117] The two-channel detection signal is multiplied with the reference signal and then passed through a low-pass filter to obtain four sets of slowly varying signals of the gyroscope:

[0118]

[0119] E = C x 2 +S x 2 +C y 2 +S y 2 = a 2 + q 2 ;

[0120] Q = 2(C x S y -C y S x ) = 2aq;

[0121]

[0122] wherein E is the total energy of the gyroscope, Q is the quadrature of the gyroscope; L is the phase-locked loop control quantity,

[0123] Since the quadrature control suppresses Q to zero, the amplitude control maintains E at the target value, the control voltage output by the quadrature control is V q , and the control voltage output by the amplitude control is V a ; since the virtual precession control loop needs to control the standing wave to precess in the form of a constant value superimposed on a sine wave, the initially given virtual precession control voltage V w0 is designed as:

[0124] V w0 = V con-0 + V sin sinλt;

[0125] V con-0 is the constant component, V sin is the amplitude of the sine component, and λ is the frequency of the sine component;

[0126] In the prior art, according to the control voltages V w0 , V a and V q , the virtual precession control force F w0 , the amplitude control force F a , and the quadrature control force F q are obtained;

[0127]

[0128] The control forces are projected onto the X / Y mode:

[0129]

[0130] where F x is the x modal control force, F y is the y modal control force, θ is the standing wave azimuth angle, F x and F y are used to act on the x and y modes of the hemispherical resonator gyroscope's resonator, respectively.

[0131] In the present application, the adjusted virtual precession control voltage V w0 is obtained in the same way as in the prior art.

[0132] While the application has been described with reference to particular embodiments, it will be understood that the examples are merely illustrative of the principles and applications of the application. It will be further understood that numerous modifications can be made to the illustrative embodiments, and that other arrangements can be devised without departing from the spirit and scope of the present application as defined by the appended claims. It will be understood that the features of the various embodiments can be combined with each other, where appropriate. It will be further understood that features described with respect to one embodiment can be used in other embodiments in combination with other features. It will be understood that features described with respect to one embodiment can be used in other embodiments in combination with other features. It will be understood that features described with respect to one embodiment can be used in other embodiments in combination with other features. It will be understood that features described with respect to one embodiment can be used in other embodiments in combination with other features. It will be understood that features described with respect to one embodiment can be used in other embodiments in combination with other features.

Claims

1. A method for online calibration of virtual precession of a hemispherical resonator gyro based on voltage fluctuation, characterized in that, The method comprises the following steps: Step 1, power on the hemispherical resonator gyroscope, make it in the initial given virtual precession control voltage The corresponding virtual precession control force Lower precession; Step 2, using the same frequency cosine reference signal output by the phase-locked loop to demodulate the vibration signal detected by the X and Y channels of the hemispherical resonator gyro in real time, to obtain the X channel cosine signal and the Y channel cosine signal ​ Step 3, for and Angle calculations are performed to obtain the standing wave azimuth. Then, the standing wave azimuth angle is... The standing wave precession velocity is obtained by performing differential calculations. ; Step 4, the initial value of the step 3 is 1, the initial value of the precession error is set to 0, the precession error is identified on line according to the standing wave precession velocity and the specific frequency reference signal by using the precession error identification algorithm. wheel recognition, The initial value of the step 4 is 1, the initial value of the precession error is set to 0, the precession error is identified on line according to the standing wave precession velocity and the specific frequency reference signal by using the precession error identification algorithm. wheel recognition, The initial value of the step 4 is 1, the initial value of the precession error is set to 0, the precession The specific frequency reference signal is a given frequency sine signal; The implementation mode of the precession error online identification algorithm comprises: Step 41, establishing a standing wave precession velocity and an initially given virtual precession control voltage between the associated models: ； wherein is a precession factor for the spherical harmonic resonator, is an input angular rate, is a detection gain, is a control gain, is a target amplitude, is a resonant frequency of the resonator, is a constant component, is an amplitude of the sinusoidal component, is a frequency of the sinusoidal component, is time; Step 42, write the correlation model into a discrete product form: ； Wherein: ； is the expected output vector for the first sampling time instant, is the standing wave precession velocity for the first sampling time instant, is the sampling time instant sequence number, , , is the observation vector for the first sampling time instant, is the time corresponding to the first sampling time instant, is the parameter vector for the first sampling time instant, is the first row first column element in the matrix, is the second row first column element in the matrix, T is the matrix transpose, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant, is the discrete quantity of the first sampling time instant. Step 43, updating the gain vector K and the covariance matrix P of the precession error identification algorithm corresponding to the m-th sampling time according to the result of step 42 . :​​ Step 44, obtain the parameter estimate of the first sampling time from the result of step 43 Step 42 and Step 42 Step 42 Step 42 Step 42 Step 42 Step 42 ； wherein ; Step 5, determine the first Check if the precession error obtained from wheel identification has converged. If yes, proceed to step 6; otherwise... Proceed to step 4; Step 6, record the initial given virtual precession control voltage and the precession error is identified, and the precession error calibration is prepared to be started; Step 7, judge whether the power-on time of the converged hemispherical resonator gyroscope reaches a time threshold value, and if yes, jump to step 8, otherwise, jump to step 6; Step 8, start precession error calibration: According to the first The precession error obtained by wheel recognition and Adjust the initial given virtual precession control voltage Make the hemispherical resonator gyro in the adjusted virtual precession control voltage The corresponding virtual precession control force under the standing wave precesses in the form of a constant superimposed specific frequency reference signal, and further makes the precession speed of the standing wave Keep constant, complete the online calibration of the virtual precession of the hemispherical resonator gyro.

2. The method according to claim 1, wherein, The value of the time threshold value is 10 minutes.

3. The method according to claim 1, wherein, Step 5, judging whether the precession error obtained by the wheel recognition converges or not The condition for judging whether the precession error obtained by the wheel recognition converges or not is: When the difference between the precession errors recognized by the two adjacent wheels is less than a set convergence threshold, it is determined that the precession errors recognized by the two adjacent wheels converge, otherwise, the precession errors recognized by the two adjacent wheels do not converge. When the difference between the precession errors recognized by the two adjacent wheels is less than a set convergence threshold, it is determined that the precession errors recognized by the two adjacent wheels converge, otherwise, the precession errors recognized by the two adjacent wheels do not converge. When the difference between the precession errors recognized by the two adjacent wheels is less than a set convergence threshold, it is determined that the precession errors recognized by the two adjacent wheels converge, otherwise, the pre 4. The method according to claim 1, wherein, Step 2 yields the X channel cosine signal and the Y channel cosine signal are implemented as ； Wherein, ； ； ； ； is a low-pass filter, and are in-phase and quadrature sinusoidal reference signals, respectively, and are the detected vibration signals of the X and Y channels, respectively, and represent the in-phase and quadrature wave amplitudes, respectively, is the initial phase of the reference signal, is the resonant frequency of the resonator, is time.

5. The method according to claim 1, wherein, In step 3, the standing wave azimuth angle is determined as .

6. The method according to claim 1, wherein, In step 43, ； ； wherein is the covariance matrix corresponding to the th sampling instant; , is the identity matrix, is a forgetting factor, selected in the range (0, 1].

7. The method according to claim 1, wherein, In step 8, the adjusted virtual precession control voltage The expression for this is: ； wherein is a constant component, is a first is a precession error of the wheel recognition, is an amplitude of the sinusoidal component, is a frequency of the sinusoidal component, is time.

8. The method according to claim 1, wherein, In step 41, the value range of the frequency f is 0 Hz to 10 Hz.

Images