Temperature drift compensation method for hemispherical resonator gyroscope

Abstract

The invention relates to a temperature drift compensation method for a hemispherical resonator gyroscope, which comprises the following steps of: designing a hemispherical resonator gyroscope temperature estimation model taking parameters such as a heater on-off state of a thermal control subsystem, heater on / off action time, cabin plate temperature, partition plate temperature and the like as input through on-orbit data analysis; a neural network model is constructed to learn the nonlinear uncertain drop process of the temperature of the hemispherical resonator gyroscope in the closed state of the heater; and a compensation model between the temperature of the hemispherical resonator gyroscope and the constant drift is established. According to the invention, the on-orbit temperature compensation problem of the hemispherical resonator gyroscope is solved, the measurement precision of the hemispherical resonator gyroscope is improved, a frequency measurement circuit and a temperature control circuit do not need to be additionally designed, and the cost is saved.

Description

Technical Field

[0001] The invention relates to a temperature drift compensation method for a hemispherical resonator gyroscope, which is suitable for temperature drift compensation of a hemispherical resonator gyroscope of a spacecraft control system and can also be extended to other similar temperature drift compensation occasions of inertial measurement units. Background Art

[0002] The hemispherical resonator gyroscope is one of the most commonly used inertial sensors on spacecraft and is an important component for achieving spacecraft attitude control. During the operation of the spacecraft control system, it is necessary to use the angular velocity information output by the hemispherical resonator gyroscope to implement angular velocity closed-loop control, or integrate the spacecraft attitude angle to determine the attitude angle. In situations with high control accuracy requirements, temperature drift compensation of the hemispherical resonator gyroscope is an indispensable task.

[0003] The temperature drift compensation methods of hemispherical resonant gyroscopes can generally be divided into two categories:

[0004] (1) Method based on frequency measurement circuit. The relationship between the resonant frequency and temperature of the hemispherical resonant gyroscope is linear. Therefore, the resonant frequency of the hemispherical resonant gyroscope can be used as an indirect measurement of the ambient temperature of the hemispherical resonator to perform temperature compensation on the gyro's scale factor and gyro zero bias. However, it is necessary to design a frequency measurement circuit with a certain resolution and sensitivity, which will increase the complexity and development cost of the single machine.

[0005] (2) Method based on temperature control circuit. By adding temperature measurement and temperature control circuits, the temperature of the gyroscope can be controlled within a certain working range or working point, and the compensation value can be designed for the specific working range or working point. However, the need to design temperature measurement and temperature control circuits will also increase the complexity and development cost of the single machine. Summary of the invention

[0006] The technology of the present invention solves the problem: overcomes the shortcomings of the prior art, proposes a temperature drift compensation method for a hemispherical resonant gyroscope, realizes the temperature drift compensation of the hemispherical resonant gyroscope when there is no direct effective measurement data on orbit, and effectively improves the accuracy of the angular velocity measurement of the hemispherical resonant gyroscope when it is impossible to introduce a frequency measurement circuit and a temperature measurement and control circuit.

[0007] The technical solution of the present invention:

[0008] A temperature drift compensation method for a hemispherical resonant gyroscope comprises the following steps:

[0009] 1) Establish a feedforward neural network for HRG temperature estimation;

[0010] 2) Using the ground test data as sample data and target data to train the feedforward neural network;

[0011] 3) During in-orbit flight, determine the switch status of the hemispherical resonant gyro heater and record the corresponding opening time t ON / Closing time t ON If the heater is off, the feedforward neural network described in step 2) is used to obtain the temperature estimate T′ of the hemispherical resonant gyroscope. HRG ; If the heater is on, the temperature estimate T′ of the HRG is obtained using the on-orbit estimation model HRG ;

[0012] 4) Using the estimated temperature value T′ of the hemispherical resonant gyroscope obtained in step 3) HRG , the scale factor and zero bias of the hemispherical resonant gyroscope are compensated in turn, and the angular velocity pulse output by the hemispherical resonant gyroscope after compensation is obtained.

[0013] Preferably, the input of the feedforward neural network in step 1) includes: the star time t corresponding to the moment when the heater is turned off OFF , the star time t corresponding to the sampling time, the cabin temperature T in the cabin where the hemispherical resonant gyroscope is located deck and the partition temperature T isolate ; The output of the feedforward neural network is the estimated value of the hemispherical resonant gyro temperature T' HRG .

[0014] Preferably, the sample data in step 2) is: the star time t corresponding to the moment when the heater is turned off OFF , the star time t corresponding to the current sampling moment, the cabin temperature T in the cabin where the hemispherical resonant gyroscope is located deck and the partition temperature T isolate .

[0015] Preferably, the target data in step 2) is the real-time measurement value T of the hemispherical resonant gyro temperature. HRG .

[0016] Preferably, the step 2) trains the feedforward neural network as follows:

[0017] 21) Let the sampling period be T s seconds, collect N groups of ground test data as sample sets, N is not less than 10, each group has M sampling points, M ≥ 100; the sampling data of each sampling point contains {T deck ,T isolaTe ,Δt,T HRG}, T HRG It indicates the gyro temperature measured in real time by the thermistor attached to the surface of the hemispherical resonant gyro during ground testing. The thermistor does not follow the hemispherical resonant gyro into orbit. Δt = tt OFF ;

[0018] 22) 70% of the data in the sample set is selected to form a training data set to train the feedforward neural network; 15% of the remaining sample set is selected to verify and test the training results of the feedforward neural network; the conditions for the end of training are:

[0019]

[0020] Among them, i is the label of the data in the training data set, and the value of i is [1, 0.7NM].

[0021] Preferably, the on-orbit estimation model in step 3) is specifically:

[0022] T′ HRG =T ON +k·(tt ON )

[0023] Among them, T ON represents the heater opening threshold temperature, which is determined by the upper-level task input; k represents the temperature rise gradient when the heater is turned on; t represents the star time corresponding to the current sampling moment; t ON Indicates the star time corresponding to the time when the heater is turned on.

[0024] Preferably, the temperature rise gradient k when the heater is turned on is in the range of 0.005 to 0.1° C. / second.

[0025] Preferably, the step 4) obtains the angular velocity pulse D output by the hemispherical resonant gyro after compensation. HRG , specifically:

[0026]

[0027] Among them, B2, B1 and B0 are zero bias compensation coefficients, which are inherent parameters of the hemispherical resonant gyroscope. The specific values ​​of them can be determined by polynomial fitting based on the laboratory test data of the hemispherical resonant gyroscope.

[0028] D″ HRG and D′ HRG are the angular velocity pulses output by the hemispherical resonant gyroscope before and after scale factor compensation, respectively. A1 and A0 are scale factor compensation coefficients, which are inherent parameters of the hemispherical resonant gyroscope. Their specific values ​​can be determined by polynomial fitting based on the laboratory test data of the hemispherical resonant gyroscope.

[0029] Compared with the prior art, the present invention has the following advantages:

[0030] 1) The present invention uses the remote measurement of non-attitude and orbit control systems such as the temperature of the entire satellite cabin panel and the partition panel to improve the performance of the control subsystem, making full use of the information flow advantages under the integrated electronic architecture.

[0031] 2) The present invention utilizes a neural network to learn the nonlinear characteristics of temperature change when the heater is turned off, thereby solving the modeling difficulties caused by uncertainty.

[0032] 3) The present invention does not rely on frequency measurement circuits and temperature control circuits, has low cost and high reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0034] In order to better describe the present invention, the present invention is described in detail below with reference to schematic diagrams and examples.

[0035] like Figure 1 As shown, the present invention provides a temperature drift compensation method for a hemispherical resonant gyroscope, comprising the following steps:

[0036] 1) Establish a feedforward neural network for hemispherical resonant gyro temperature estimation and determine the specific structure of the feedforward neural network;

[0037] The HRG is usually installed in a cabin in the service module of a spacecraft, which is a closed space divided by the cabin board and the partition board. When the heater is turned off, the temperature drop rate of the HRG is mainly affected by the cabin board temperature and the partition board temperature where the HRG is installed. Therefore, the input of the feedforward neural network includes: the star time t corresponding to the moment when the heater is turned off OFF , the star time t corresponding to the sampling time, the cabin temperature T in the cabin where the hemispherical resonant gyroscope is located deck and the partition temperature T isolate ; The output of the feedforward neural network is the estimated value of the hemispherical resonant gyro temperature T' HRG .

[0038] 2) Using the ground test data as sample data and target data, the parameters of each layer of the feedforward neural network are trained;

[0039] Step 2) The sample data used in the training is: the star time t corresponding to the heater off time OFF , the star time t corresponding to the current sampling moment, the cabin temperature T in the cabin where the hemispherical resonant gyroscope is located deck and the partition temperature T isolate .

[0040] Step 2) The target data used in the training is the real-time measurement value of the hemispherical resonant gyro temperature T HRGBy training the feedforward neural network, the feedforward neural network can learn the nonlinear uncertainty relationship between the real-time measurement values ​​of the satellite cabin temperature, the satellite partition temperature and the hemispherical resonant gyro temperature when the heater is turned off, so that the feedforward neural network has an acceptable temperature estimation capability, and it is considered that a feedforward neural network that can be used in orbit has been obtained;

[0041] 3) During in-orbit flight, determine the switch status of the hemispherical resonant gyro heater and record the corresponding opening time t ON / Closing time t ON If the heater is off, the feedforward neural network described in step 2) is used to obtain the temperature estimate T′ of the hemispherical resonant gyroscope. HRG ; If the heater is on, the temperature estimate T′ of the HRG is obtained using the on-orbit estimation model HRG ;

[0042] The on-orbit estimation model in step 3) is specifically:

[0043] T′ HRG =T ON +k·(tt ON )

[0044] Among them, T ON represents the heater start-up threshold temperature, which is determined by the upper-level task input; k represents the temperature rise gradient when the heater is turned on, and the value range is 0.005~0.1℃ / s; t represents the star time corresponding to the current sampling moment; t ON Indicates the star time corresponding to the time when the heater is turned on.

[0045] 4) Using the estimated temperature value T′ of the hemispherical resonant gyroscope obtained in step 3) HRG As input parameters, the scale factor and zero bias of the hemispherical resonant gyroscope are compensated in turn to obtain the angular velocity pulse output by the hemispherical resonant gyroscope after compensation.

[0046] The feedforward neural network designed in step 1) is specifically as follows:

[0047] 11) The input layer is designed with three neurons, x = (x1, x2, x3), and meets the following conditions: x1 is the temperature T of the entire satellite cabin in the cabin where the hemispherical resonant gyro is located. deck ; x2 is taken as the temperature T of the entire satellite partition in the cabin where the hemispherical resonant gyro is located isolate ; x3 is taken as the satellite time difference between the current sampling time and the heater shut-off time, that is, x3 = Δt = tt OFF , let u i represents the output of the i-th neuron in the input layer, then:

[0048] u i=x i ,,i∈[1,3]

[0049] 12) Design a single hidden layer, which contains 10 neurons, and the output h of the jth neuron is j for:

[0050]

[0051] Among them, f j (·) is the transmission function of the hidden layer neurons, h j represents the output of the jth neuron in the hidden layer; ω i,j is the connection weight between the i-th neuron in the input layer and the j-th neuron in the hidden layer, and its value range is [0,1]; b j is the jth bias value, the initial value can be any small real number, and the value range is generally 0.01 to 1;

[0052] 13) Design the output layer as a single neuron, and the output y is the estimated value T′ of the hemispherical resonant gyroscope temperature HRG , which is calculated as follows:

[0053]

[0054] Among them, ω j is the connection weight between the jth neuron in the hidden layer and the neuron in the output layer, and its value range is [0,1]; b is the bias value of the output layer, and its initial value can be any small real number, and its value range is generally 0.01~1.

[0055] The step 2) trains the parameters of each layer of the feedforward neural network, as follows:

[0056] 21) Let the sampling period be T s seconds, collect N groups (N is not less than 10) (each group has M sampling points, M ≥ 100) of ground test data as the sample set, and the sampling data of each sampling point contains {T deck ,T isolate ,Δt,T HRG}, T HRG It indicates the gyro temperature measured in real time by the thermistor attached to the surface of the hemispherical resonant gyro during ground testing. The thermistor does not follow the hemispherical resonant gyro into orbit. In one embodiment of the present invention, the sampling period is 8 seconds, and 36 groups of samples are collected, each with 1000 sampling points.

[0057] 22) 70% of the data in the sample set is selected to form a training data set to train the neural network. Another 15% is taken from the remaining sample set for verification and testing of the neural network training results. The neural network training method is a well-known technology. The conditions for the end of training are:

[0058]

[0059] That is, the sum of squares of the error between the estimated value and the measured value is less than 10 -5 i is the label of the data in the training data set, and the value of i is [1, 0.7NM].

[0060] The step 4) of the hemispherical resonant gyro scale factor compensation model and the zero bias compensation model is as follows:

[0061] 41) The scale factor compensation model of the hemispherical resonant gyro is:

[0062]

[0063] Where: D″ HRG and D′ HRG are the angular velocity pulses output by the hemispherical resonant gyroscope before and after scale factor compensation, respectively. A1 and A0 are scale factor compensation coefficients, which are inherent parameters of the hemispherical resonant gyroscope. Their specific values ​​can be determined by polynomial fitting based on the laboratory test data of the hemispherical resonant gyroscope.

[0064] 42) The zero bias compensation model of the hemispherical resonant gyro is:

[0065]

[0066] Where: D GRG is the angular velocity pulse output by the hemispherical resonant gyro after zero bias compensation, that is, the angular velocity pulse output by the hemispherical resonant gyro after compensation, B2, B1 and B0 are zero bias compensation coefficients, which are inherent parameters of the hemispherical resonant gyro. Their specific values ​​can be determined by polynomial fitting based on the laboratory test data of the hemispherical resonant gyro.

[0067] The corresponding values ​​of the hemispherical resonant gyroscope according to an embodiment of the present invention are A1=1.4552, A0=19991.31, B2=0.0000048, B1=0.00057, and B0=0.00061.

[0068] The present invention is aimed at a type of hemispherical resonant gyroscope that is not designed with a frequency measurement circuit and a temperature measurement and control circuit, or a hemispherical resonant gyroscope whose frequency measurement circuit (temperature measurement and control circuit) is abnormal on-orbit. It comprehensively uses prior knowledge and the learning ability of a feedforward neural network to construct an on-orbit estimation model for the temperature of the hemispherical resonant gyroscope, and comprehensively uses on-orbit data and laboratory data to identify model parameters and compensation parameters. It realizes the temperature drift compensation of the hemispherical resonant gyroscope in the case of no direct effective measurement data on-orbit, and effectively improves the accuracy of the angular velocity measurement of the hemispherical resonant gyroscope when the frequency measurement circuit and the temperature measurement and control circuit cannot be introduced.

[0069] Although the present invention has been disclosed as above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solution of the present invention by using the above disclosed methods and technical contents without departing from the spirit and scope of the present invention. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention shall belong to the protection scope of the technical solution of the present invention. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0070] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

Claims

1. A temperature drift compensation method for a hemispherical resonant gyroscope, characterized in that: The steps include: 1) Establish a feedforward neural network for HRG temperature estimation; 2) Using the ground test data as sample data and target data to train the feedforward neural network; 3) When flying on orbit, determine the switch state of the hemispherical resonant gyro heater and record the corresponding opening and closing times; if the heater is in the off state, use the feedforward neural network described in step 2) to obtain the temperature estimate T′ of the hemispherical resonant gyro HRG ; If the heater is on, the temperature estimate T′ of the HRG is obtained using the on-orbit estimation model HRG ; 4) Using the estimated temperature value T′ of the hemispherical resonant gyroscope obtained in step 3) HRG , the scale factor and zero bias of the hemispherical resonant gyroscope are compensated in turn, and the angular velocity pulse output by the hemispherical resonant gyroscope after compensation is obtained.

2. A temperature drift compensation method for a hemispherical resonant gyroscope according to claim 1, characterized in that: Step 1) The input of the feedforward neural network includes: the star time t corresponding to the moment when the heater is turned off OFF , the star time t corresponding to the sampling time, the cabin temperature T in the cabin where the hemispherical resonant gyroscope is located deck and the partition temperature T isolate ; The output of the feedforward neural network is the estimated value of the hemispherical resonant gyro temperature T' HRG .

3. A temperature drift compensation method for a hemispherical resonant gyroscope according to claim 2, characterized in that: Step 2) The sample data is: the star time t corresponding to the heater off time OFF , the star time t corresponding to the current sampling moment, the cabin temperature T in the cabin where the hemispherical resonant gyroscope is located deck and the partition temperature T isolate .

4. A temperature drift compensation method for a hemispherical resonant gyroscope according to claim 3, characterized in that: Step 2) The target data is the real-time measurement value T of the hemispherical resonant gyro temperature. HRG .

5. A temperature drift compensation method for a hemispherical resonant gyroscope according to claim 4, characterized in that: The step 2) trains the feedforward neural network as follows: 21) Let the sampling period be T s seconds, collect N groups of ground test data as sample sets, N is not less than 10, each group has M sampling points, M ≥ 100; the sampling data of each sampling point contains {T deck ,T isolate ,Δt,T HRG }, T HRG It indicates the gyro temperature measured in real time by the thermistor attached to the surface of the hemispherical resonant gyro during ground testing. The thermistor does not follow the hemispherical resonant gyro into orbit. Δt = tt OFF ; 22) 70% of the data in the sample set is selected to form a training data set to train the feedforward neural network; 15% of the remaining sample set is selected to verify and test the training results of the feedforward neural network; the conditions for the end of training are: Among them, i is the label of the data in the training data set, and the value of i is [1, 0.7NM].

6. A temperature drift compensation method for a hemispherical resonant gyroscope according to any one of claims 1 to 5, characterized in that: The on-orbit estimation model in step 3) is specifically: T′ HRG =T ON +k·(t-t ON ) Among them, T ON represents the heater opening threshold temperature, which is determined by the upper-level task input; k represents the temperature rise gradient when the heater is turned on; t represents the star time corresponding to the current sampling moment; t ON Indicates the star time corresponding to the time when the heater is turned on.

7. A temperature drift compensation method for a hemispherical resonant gyroscope according to claim 6, characterized in that: The temperature rise gradient k when the heater is turned on ranges from 0.005 to 0.1°C / second.

8. The temperature drift compensation method for a hemispherical resonant gyroscope according to claim 6, characterized in that: The step 4) obtains the angular velocity pulse D output by the hemispherical resonant gyro after compensation. HRG , specifically: Among them, B2, B1 and B0 are zero bias compensation coefficients, which are inherent parameters of the hemispherical resonant gyroscope. The specific values ​​of them can be determined by polynomial fitting based on the laboratory test data of the hemispherical resonant gyroscope. D″ HRG and D′ HRG are the angular velocity pulses output by the hemispherical resonant gyroscope before and after scale factor compensation, respectively. A1 and A0 are scale factor compensation coefficients, which are inherent parameters of the hemispherical resonant gyroscope. Their specific values ​​can be determined by polynomial fitting based on the laboratory test data of the hemispherical resonant gyroscope.

Images