An atomic interference gyroscope data update frequency intelligent promotion method and system

Abstract

The application discloses an atomic interference gyroscope data update frequency intelligent promotion method and system, belongs to the field of inertial measurement and navigation positioning, and solves the problem that the data update rate and the measurement accuracy of the atomic interference gyroscope influence each other. The application collects data by using the atomic interference gyroscope, combines an IGCSMK GPR prediction model based on a sparsification improved GCSMK-GPR prediction model, realizes promotion of the data update frequency, then collects information by using a laser gyroscope, monitors and corrects output data of the atomic interference gyroscope, modifies model parameters in the sparsification GCSMK-GPR prediction algorithm, and uses the model parameters as parameters for prediction at the next moment, so that the prediction accuracy is improved. The application is suitable for promotion of the atomic interference gyroscope data update frequency and measurement of the accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of inertial measurement and navigation positioning, and particularly to the field of atomic interference gyro data updating. BACKGROUND

[0002] As a kind of full autonomous navigation and positioning technology, inertial navigation has the advantages of all-weather, good concealment, not easy to be disturbed, strong survivability, etc., and is widely used in navigation and positioning in various fields. Inertial navigation system (INS) relies on inertial measurement components (gyroscopes and accelerometers) to real-time sense the angular motion and linear motion state of the carrier, and outputs the navigation information such as the velocity, position and attitude of the carrier through navigation calculation. Among them, the gyroscope as the core sensor of the inertial navigation system, its precision directly affects the performance of the inertial navigation system.

[0003] In recent decades, under the impetus of the urgent needs of national defense construction, high-precision gyroscope technology has developed rapidly, from the first generation of rotor gyroscope based on Newtonian mechanics to the second generation of optical gyroscope based on wave optics. Among them, the electrostatic gyroscope as the highest precision rotor gyroscope, the highest precision can reach 10 -6 degree / hour, is widely equipped in strategic nuclear submarines, but the rotor drift caused by the process limitations of beryllium rotor and the problems such as rotor speed decay and long stabilization time seriously affect the dynamic measurement precision of the electrostatic gyroscope, and restrict the further development of the electrostatic gyroscope inertial navigation system; as a representative of optical gyroscopes, the measurement precision of fiber optic gyroscopes and laser gyroscopes can reach 10 -4 degree / hour, and are increasingly applied to tactical weapons and various carriers, however, in the actual complex environment, affected by multiple external disturbance coupling factors, the internal stress field of the fiber optic gyroscope changes, leading to unstable dynamic scale factor, the laser gyroscope itself is coupled with external disturbance, making the shock absorber structure modal change, and it is difficult to optimize the shock absorbing mode, which cannot guarantee the high precision measurement in the actual dynamic environment, and is difficult to meet the needs of future high-precision inertial navigation systems.

[0004] With the birth of three Nobel Physics Prizes in the field of quantum, atomic excitation, atomic capture, quantum state superposition and atomic group temperature reduction technologies have developed rapidly, and atomic gyro inertial navigation system based on quantum mechanics has become a research hotspot in the navigation field abroad. Among them, the atomic interference gyroscope based on the principle of atomic de Broglie wave interference is currently the gyroscope with the highest precision potential, and the theoretical measurement precision can reach 10 -10Degrees per hour is becoming a research focus for ultra-high-precision inertial navigation systems. The U.S. Defense Advanced Research Projects Agency (DARPA) considers atomic inertial sensing technology, centered around atomic interferometry, to be the next generation of dominant inertial technology, and is pursuing the realization of ultra-high-precision military inertial navigation systems with autonomous positioning accuracy of 5 meters per hour.

[0005] Although atomic interferometer gyroscopes and related technologies have achieved certain breakthroughs, the technical bottleneck of low data update rates remains unresolved, making their performance insufficient for navigation applications in dynamic environments. To this end, numerous research institutions at home and abroad have conducted extensive research, with representative teams including the French National Laboratory for Metrology and Testing (LNE-SYRTE) and the Wuhan Institute of Physics and Mathematics of the Chinese Academy of Sciences. In early 2016, the LNE-SYRTE team leveraged the advantages of the four-pulse configuration for rotational speed measurement and designed a new continuous measurement configuration. This eliminates the dead time caused by atom preparation and detection, aligning the operating cycle with the measurement cycle and achieving maximum rotational measurement sensitivity at the same data update rate. Subsequently, the team proposed a three-pulse configuration for the parabolic gyroscope, which separates common-mode and non-common-mode errors, maximizing the data update frequency while maintaining measurement accuracy. In 2013, the Wuhan Institute of Physics and Mathematics of the Chinese Academy of Sciences proposed an overlapping interferometer scheme, which eliminates the impact of dead time caused by the preparation and collection of atomic clusters. While increasing the data rate, it also maximizes the sensitivity advantage of the time-domain atom interferometer. It is currently a more practical solution for improving data update rates.

[0006] However, during the implementation of the above solution, it was found that the above solution still has some defects:

[0007] From the working principle of the atomic interferometer gyroscope, we can know that higher measurement accuracy means better atomic controllability, which requires longer atomic preparation and cooling time, which corresponds to a longer data cycle. A significant increase in the data update rate will inevitably affect the free flight time of the atoms, which in turn affects the measurement accuracy. Therefore, the data update frequency and measurement accuracy of the atomic interferometer gyroscope are two conflicting technical indicators. From the current technological status, it is extremely difficult to improve the data update frequency of the atomic interferometer gyroscope at the device level. Summary of the Invention

[0008] The present invention provides a method and system for intelligently improving the data update frequency of an atomic interferometer gyroscope, which solves the problems of low data update rate of the atomic interferometer gyroscope, difficulty in meeting the navigation application requirements in actual dynamic environments, and mutual influence between the data update rate and measurement accuracy of the atomic interferometer gyroscope.

[0009] An atomic interference gyroscope data update frequency intelligent promotion system, the system comprises:

[0010] An atomic interference gyroscope, a laser gyroscope and a vibration isolation platform;

[0011] The atomic interference gyroscope and the laser gyroscope are both fixed on the vibration isolation platform;

[0012] An intelligent control unit is embedded in the atomic interference gyroscope;

[0013] The atomic interference gyroscope and the laser gyroscope collect data and send the data to the intelligent control unit;

[0014] The intelligent control unit comprises a sparse GCSMK-GPR predictor unit and a sparse GCSMK-GPR-based atomic interference gyroscope data update frequency promotion subunit;

[0015] The sparse GCSMK-GPR-based atomic interference gyroscope data update frequency promotion subunit is used for detecting whether the atomic interference gyroscope has an output, if there is an output, the output of the atomic interference gyroscope is output data, if there is no output, the sparse GCSMK-GPR predictor unit is used for prediction to obtain output data;

[0016] The sparse GCSMK-GPR-based atomic interference gyroscope data update frequency promotion subunit is also used for consistency detection and correction of the output data and modification thereof.

[0017] Further, the X axis, the Y axis and the Z axis of the laser gyroscope are respectively parallel to the X axis, the Y axis and the Z axis of the atomic interference gyroscope.

[0018] Further, the system further comprises a carrier or a mounting base.

[0019] Further, the vibration isolation platform is fixed on the carrier or the mounting base.

[0020] Further, the atomic interference gyroscope and the laser gyroscope are not more than 10 cm away.

[0021] The present application also provides an atomic interference gyroscope data update frequency intelligent promotion method, the method is realized based on the following device, the device comprises an atomic interference gyroscope, a laser gyroscope and a vibration isolation platform;

[0022] The atomic interference gyroscope and the laser gyroscope are both fixed on the vibration isolation platform;

[0023] The atomic interference gyroscope and the laser gyroscope collect data and send the data to an intelligent control unit;

[0024] The method comprises:

[0025] Collect and store data sent by the atomic interference gyroscope;

[0026] A general spectrum mixed component kernel function Gaussian process regression prediction model is established, and the prediction model is subjected to sparse processing;

[0027] Optimization is performed on parameters of the model according to the stored data, and optimal model parameters are obtained;

[0028] It is detected whether the atomic interference gyroscope has an output, if the atomic interference gyroscope has an output, the output of the atomic interference gyroscope is output data, if the atomic interference gyroscope has no output, the optimal model parameters are used for prediction to obtain output data;

[0029] Data are collected by a laser gyroscope, consistency detection and correction are performed on the output data according to the collected data, and the output data are corrected, and the atomic interference gyroscope data update frequency intelligent improvement method is completed.

[0030] Further, the data collected by the atomic interference gyroscope are long-term data in a dynamic environment.

[0031] Further, the data stored by the atomic interference gyroscope are historical data.

[0032] Further, the parameters of the model are kernel functions.

[0033] Further, the data collected by the laser gyroscope are dynamic information.

[0034] The present application has the following beneficial effects:

[0035] After the atomic interference gyroscope data in a dynamic environment are analyzed by using the general spectrum mixed component kernel function Gaussian process regression model, the present application can obtain the mutual dependence relationship between different base components of the atomic interference gyroscope data in the dynamic environment, and the sparse processing can further reduce the time complexity of the algorithm, so that the data in the sampling interval of the atomic interference gyroscope can be accurately short-term predicted by using the model, and the accuracy of the prediction can be further improved after the correction of the laser gyroscope data is introduced. The present application can improve the data update frequency to a certain extent without changing the performance of the atomic interference gyroscope device, and can ensure the measurement accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1Figure 1 is a schematic diagram of a device position for implementing the atomic interference gyroscope data update frequency intelligent promotion scheme of Embodiment One and Embodiment Two, wherein 1 is an atomic interference gyroscope, 2 is a laser gyroscope, and 3 is a vibration isolation platform.

[0037] Figure 2 Figure 1 is a schematic diagram of a device position for implementing the atomic interference gyroscope data update frequency intelligent promotion scheme of Embodiment One and Embodiment Two, wherein 1 is an atomic interference gyroscope, 2 is a laser gyroscope, and 3 is a vibration isolation platform.

[0038] Figure 3 Figure 1 is a schematic diagram of a device position for implementing the atomic interference gyroscope data update frequency intelligent promotion scheme of Embodiment One and Embodiment Two, wherein 1 is an atomic interference gyroscope, 2 is a laser gyroscope, and 3 is a vibration isolation platform. DETAILED DESCRIPTION

[0039] Embodiment One: Refer to Figure 1 、 Figure 2 、 Figure 3 This embodiment is described.

[0040] An atomic interference gyroscope data update frequency intelligent promotion system, the system comprising:

[0041] An atomic interference gyroscope, a laser gyroscope, and a vibration isolation platform.

[0042] The atomic interference gyroscope and the laser gyroscope are both fixed on the vibration isolation platform.

[0043] An intelligent control unit is embedded in the atomic interference gyroscope.

[0044] The atomic interference gyroscope and the laser gyroscope collect data and send the data to the intelligent control unit.

[0045] The intelligent control unit comprises a sparse GCSMK-GPR prediction subunit and an atomic interference gyroscope data update frequency promotion subunit based on sparse GCSMK-GPR.

[0046] The atomic interference gyroscope data update frequency promotion subunit based on sparse GCSMK-GPR is used to detect whether the atomic interference gyroscope has an output. If there is an output, the output of the atomic interference gyroscope is output data. If there is no output, the sparse GCSMK-GPR prediction subunit makes a prediction to obtain output data.

[0047] The atomic interference gyroscope data update frequency promotion subunit based on sparse GCSMK-GPR is also used to detect and correct the consistency of the output data and to correct it.

[0048] Specifically:

[0049] As shown in Figure 2 and Figure 3 ,

[0050] First, the data output by the atomic interference gyroscope AIG inertial sensor is collected. If the AIG has output, the output of the AIG is the output data. If the AIG has no output, first, the kernel function form of the general convolution spectrum mixed component kernel function GCSMK is determined, the initial value of the hyperparameter is set, and the prior model of the Gaussian process regression GPR is set. The optimal hyperparameter is obtained by using the historical data stored by the atomic interference gyroscope. The Gaussian process regression prediction model, that is, the sparse GCSMK-GPR prediction model IGCSMK GPR, is established according to the prior model and the optimal hyperparameter. The output of the AIG inertial sensor is predicted by using the IGCSMK GPR prediction model, and the output data is obtained.

[0051] The dynamic information is collected and preprocessed by using the laser gyroscope. The output data is monitored and corrected by using the dynamic information. The model parameters in the sparse GCSMK-GPR prediction algorithm are modified as the parameters for prediction at the next moment, so as to improve the prediction accuracy.

[0052] From the perspective of the device, the data update frequency and the measurement accuracy of the atomic interference gyroscope are two conflicting technical indicators. If the measurement accuracy is to be improved, the data update frequency will be inevitably lost. However, the present method processes the data output by the atomic interference gyroscope to improve the update frequency without changing the performance of the device.

[0053] The consistency monitoring method is as follows: according to the historical output data of the laser gyroscope and the atomic interference gyroscope, the variances of the output data of the two are obtained, the predicted value deviation of the atomic interference gyroscope is obtained by subtracting the output data of the two, and the threshold is set according to the size of the variance. If the predicted value deviation of the atomic interference gyroscope is greater than the threshold, it is considered that the prediction error is too large, and the predicted value is inaccurate. Therefore, the output data of the laser gyroscope is used for correction. The method is as follows: if the output standard deviations of the two are σ1 and σ2, and the outputs are x1 and x2, the corrected output value is: Since the laser gyroscope data is introduced to correct the prediction error, the prediction accuracy is improved to some extent.

[0054] The purpose of using the consistency monitoring and correction algorithm is to correct the output data of the atomic interference gyroscope by using the output data of the laser gyroscope, and then store the corrected output data of the atomic interference gyroscope in the historical database. Then, in the process of optimizing the hyperparameter, the corrected historical data corrects the hyperparameter, so as to improve the accuracy of the prediction model of the atomic interference gyroscope.

[0055] Since the laser gyroscope data is introduced to correct the prediction error, the prediction accuracy is improved to some extent.

[0056] Embodiment two: refer to Figure 1The embodiment is illustrated.

[0057] The embodiment is a further illustration of the atomic interference gyroscope and the laser gyroscope in the atomic interference gyroscope data update frequency intelligent promotion system of the first embodiment.

[0058] The X-axis, Y-axis and Z-axis of the laser gyroscope are parallel to the X-axis, Y-axis and Z-axis of the atomic interference gyroscope respectively.

[0059] Specifically,

[0060] The parallelism between the three axes of the two sets of gyroscopes can avoid the influence of the installation deviation angle between the two sets of gyroscopes on the difference between the output angular velocities of the two sets of gyroscopes.

[0061] If the above requirements cannot be met, the two sets of gyroscopes can be fixedly installed on a conversion board, and then the installation error angle of the two sets of gyroscopes is calibrated, and then the influence is suppressed.

[0062] The third embodiment

[0063] The embodiment is a further illustration of the atomic interference gyroscope data update frequency intelligent promotion system of the first embodiment or the second embodiment.

[0064] The system also includes a carrier or a mounting base.

[0065] The fourth embodiment

[0066] The embodiment is a further illustration of the vibration isolation platform in the atomic interference gyroscope data update frequency intelligent promotion system of the third embodiment.

[0067] The vibration isolation platform is fixed on the carrier or the mounting base.

[0068] Specifically,

[0069] The gyro is sensitive to the angular motion of the carrier, so the gyro, that is, the system, needs to be installed on the carrier, and the motion of the carrier can make the gyro sensitive to dynamic information.

[0070] The system can also be installed on the mounting base, and then the mounting base is installed on the carrier or the turntable, or other objects to be measured.

[0071] The fifth embodiment

[0072] The embodiment is a further illustration of the atomic interference gyroscope and the laser gyroscope in the atomic interference gyroscope data update frequency intelligent promotion system of the first embodiment.

[0073] The atomic interference gyroscope in the embodiment is not more than 10 cm away from the laser gyroscope.

[0074] Specifically,

[0075] The close position is to avoid the influence of the installation platform deformation and other factors on the difference between the output angular velocities of the two sets of gyroscopes.

[0076] Embodiment six:

[0077] The atomic interference gyroscope data update frequency intelligent promotion method in the embodiment is realized based on the following device, which comprises an atomic interference gyroscope, a laser gyroscope and a vibration isolation platform.

[0078] The atomic interference gyroscope and the laser gyroscope are both fixed on the vibration isolation platform.

[0079] The atomic interference gyroscope and the laser gyroscope collect data and send the data to an intelligent control unit.

[0080] The method comprises:

[0081] Collect and store the data sent by the atomic interference gyroscope;

[0082] A general spectrum mixed component kernel function Gaussian process regression prediction model is established, and the prediction model is sparsified.

[0083] The parameters of the model are optimized according to the stored data, and the optimal model parameters are obtained.

[0084] It is detected whether the atomic interference gyroscope has an output, if the atomic interference gyroscope has an output, the output of the atomic interference gyroscope is output data, if the atomic interference gyroscope has no output, the optimal model parameters are used for prediction to obtain output data.

[0085] The laser gyroscope is used to collect data, the output data is detected and corrected according to the collected data, and the output data is corrected, so as to complete the atomic interference gyroscope data update frequency intelligent promotion method.

[0086] Specifically,

[0087] The method comprises:

[0088] The atomic interference gyroscope is used to collect long-term data in a dynamic environment, and historical data is stored.

[0089] A general spectrum mixed component kernel function Gaussian process regression prediction model is established, and the prediction model is sparsified.

[0090] According to the data, the parameters of the model, that is, the kernel function, are optimized, and optimal model parameters are obtained;

[0091] Detecting whether the atomic interference gyroscope has an output, if the atomic interference gyroscope has an output, the output of the atomic interference gyroscope is output data, if the atomic interference gyroscope has no output, the hyperparameters are short-term predicted by using the sparse prediction model, and output data is obtained;

[0092] The laser gyroscope is used to collect dynamic information, the output data is detected and corrected for consistency according to the information, and the hyperparameters are corrected, so as to improve the accuracy of the prediction model of the atomic interference gyroscope, and complete the atomic interference gyroscope data update frequency intelligent improvement method.

[0093] Embodiment seven:

[0094] The embodiment is a further illustration of the data collected by the atomic interference gyroscope in the atomic interference gyroscope data update frequency intelligent improvement method of embodiment six.

[0095] The data collected by the atomic interference gyroscope in the embodiment is long-term data in a dynamic environment.

[0096] Embodiment eight:

[0097] The embodiment is a further illustration of the data collected by the atomic interference gyroscope in the atomic interference gyroscope data update frequency intelligent improvement method of embodiment six.

[0098] The data collected by the atomic interference gyroscope in the embodiment is historical data.

[0099] Embodiment nine:

[0100] The embodiment is a further illustration of the parameters of the model in the atomic interference gyroscope data update frequency intelligent improvement method of embodiment six.

[0101] The parameters of the model in the embodiment are kernel functions.

[0102] Specifically:

[0103] When GPR training is performed, the kernel function needs to be set first. Common kernel functions include square exponential covariance and quadratic rational covariance. For example, the square exponential kernel function expression is as follows:

[0104]

[0105] Then the kernel parameter can be written as Where r=x i -x jrepresents the Euclidean distance between two input sample points; l is defined as a feature length scale parameter, and the greater the value, the smaller the correlation between the input points; is the variance of the sample.

[0106] Embodiment Ten

[0107] The embodiment is a further illustration of the information collected by the laser gyroscope in the atomic interference gyroscope data update frequency intelligent promotion method of embodiment six.

[0108] The information collected by the laser gyroscope in the embodiment is dynamic information.

Claims

1. An atomic interferometer gyroscope data update frequency intelligent improvement system, characterized in that: The system comprises: Atomic interferometer gyroscopes, laser gyroscopes, and vibration isolation platforms; The atomic interferometer gyroscope and the laser gyroscope are both fixed on the vibration isolation platform; An intelligent control unit is embedded in the atomic interferometer gyroscope; The atomic interferometer gyroscope and the laser gyroscope collect data and send the data to the intelligent control unit; The intelligent control unit includes a sparse GCSMK-GPR prediction subunit and an atomic interferometer gyroscope data update frequency improvement subunit based on the sparse GCSMK-GPR; The atomic interferometer gyroscope data update frequency improvement subunit based on sparse GCSMK-GPR is used to detect whether the atomic interferometer gyroscope has output. If there is output, the output of the atomic interferometer gyroscope is output data. If there is no output, the sparse GCSMK-GPR prediction subunit performs prediction to obtain output data. The atomic interferometer gyro data update frequency improvement subunit based on sparse GCSMK-GPR is also used to perform consistency detection and correction on the output data, and to modify it; The X-axis, Y-axis and Z-axis of the laser gyroscope are respectively parallel to the X-axis, Y-axis and Z-axis of the atomic interferometer gyroscope; The distance between the atomic interferometer gyroscope and the laser gyroscope is no more than 10 cm; The consistency detection and correction adopts the following method: Based on the historical output data of the laser gyro and the atomic interferometer gyroscope, the variance of the output data of the two is obtained respectively. The difference between the two output data is used to obtain the predicted value deviation of the atomic interferometer gyroscope. A threshold is set according to the size of the variance. If the deviation of the predicted value of the atomic interferometer gyroscope is greater than the threshold, the predicted value is inaccurate and is corrected using the output data of the laser gyroscope: If the standard deviation of the two outputs is , the output of both is , then the corrected output value is: ； The corrected output values ​​are used to modify the model parameters in the sparse GCSMK-GPR prediction algorithm and serve as the parameters for prediction at the next moment, thereby improving the prediction accuracy.

2. The system for intelligently improving the data update frequency of an atomic interferometer gyroscope according to claim 1, characterized in that: The system also includes a carrier or mounting base.

3. The system for intelligently improving the data update frequency of an atomic interferometer gyroscope according to claim 2, characterized in that: The vibration isolation platform is fixed on the carrier or the mounting base.

4. A method for intelligently improving the data update frequency of an atomic interferometer gyroscope, characterized in that: The method is implemented based on the following device, which includes an atomic interferometer gyroscope, a laser gyroscope and a vibration isolation platform; The atomic interferometer gyroscope and the laser gyroscope are both fixed on the vibration isolation platform; The atomic interferometer gyroscope and the laser gyroscope collect data and send the data to the intelligent control unit; The method comprises: Collect and store data sent by atomic interferometer gyroscope; Establishing a general spectral mixture component kernel function Gaussian process regression prediction model, and performing a sparse processing on the prediction model; Optimizing the parameters of the model according to the stored data to obtain the optimal model parameters; detecting whether the atomic interferometer gyroscope has an output; if the atomic interferometer gyroscope has an output, the output of the atomic interferometer gyroscope is output data; and if the atomic interferometer gyroscope has no output, performing prediction using the optimal model parameters to obtain output data; Using a laser gyroscope to collect data, performing consistency detection and correction on the output data based on the collected data, and correcting the data, thereby completing the method for intelligently improving the data update frequency of the atomic interferometer gyroscope; The consistency detection and correction adopts the following method: Based on the historical output data of the laser gyro and the atomic interferometer gyroscope, the variance of the output data of the two is obtained respectively. The difference between the two output data is used to obtain the predicted value deviation of the atomic interferometer gyroscope. A threshold is set according to the size of the variance. If the deviation of the predicted value of the atomic interferometer gyroscope is greater than the threshold, the predicted value is inaccurate and is corrected using the output data of the laser gyroscope: If the standard deviation of the two outputs is , the output of both is , then the corrected output value is: ； The corrected output values ​​are used to modify the model parameters in the sparse GCSMK-GPR prediction algorithm and serve as the parameters for prediction at the next moment, thereby improving the prediction accuracy.

5. The method for intelligently improving the data update frequency of an atomic interferometer gyroscope according to claim 4, characterized in that: The data collected by the atomic interferometer gyroscope is long-term data in a dynamic environment.

6. The method for intelligently improving the data update frequency of an atomic interferometer gyroscope according to claim 4, characterized in that: The data stored in the atomic interferometer gyroscope are historical data.

7. The method for intelligently improving the data update frequency of an atomic interferometer gyroscope according to claim 4, characterized in that: The parameters of the model are kernel functions.

8. The method for intelligently improving the data update frequency of an atomic interferometer gyroscope according to claim 4, characterized in that: The data collected by the laser gyroscope is dynamic information.

Images