A calibration method and device for an inertial navigation system based on a grating sensor

Abstract

The present invention discloses a calibration method and device for an inertial navigation system based on a grating sensor. The method includes: dynamically adjusting the sampling frequencies and data acquisition strategies of the grating sensor and the IMU; dynamically adjusting the sensor error correction weights according to the motion state and environmental changes to optimize the error compensation effect and provide an error correction strategy for calibration; using an adaptive partial differential equation algorithm to real-time correct the partial differential equations of the error model and the adaptive correction coefficients and dynamically adjust the error compensation strategy; introducing a virtual feedback correction loop mechanism and a deep reinforcement learning strategy, and using a virtual environment model and real-time feedback signals to optimize the calibration behavior, dynamically adjust the sensor weights, and achieve adaptive calibration; adopting an incremental self-learning calibration mechanism, and performing adaptive incremental optimization based on the error accumulation monitoring during operation and historical data to dynamically adjust the calibration parameters. The present invention breaks through the limitations of static calibration and can achieve high-precision error compensation in a complex dynamic environment.

Description

Technical Field

[0001] The present invention relates to the technical field of navigation, and particularly to a calibration method and device for an inertial navigation system based on a grating sensor. Background Art

[0002] As an autonomous positioning and navigation technology, the Inertial Navigation System (INS) has a wide range of applications in the fields of aviation, aerospace, unmanned driving, industrial measurement, etc. INS mainly relies on accelerometers and gyroscopes in the Inertial Measurement Unit (IMU) to calculate attitude, velocity, and position information. However, due to the inherent errors in the IMU, such as zero-bias drift, scale factor error, installation error, etc., these errors will accumulate continuously during long-term operation, resulting in a decrease in navigation accuracy. Therefore, high-precision calibration and error compensation of INS are the keys to improving the system stability and accuracy.

[0003] As a high-precision displacement measurement device, the grating sensor has been introduced into the error correction process of the inertial navigation system in recent years due to its advantages of strong anti-interference ability, high precision, and good long-term stability. The grating sensor can provide high-resolution displacement information and provide a reference benchmark for the error calibration of the IMU. By fusing the grating sensor and inertial navigation data, the cumulative error of the IMU can be effectively reduced, and the accuracy of the system can be improved. However, how to achieve high-precision fusion of the two in a complex environment is still a challenging problem.

[0004] Currently, traditional calibration methods for inertial navigation systems mainly use static or quasi-static calibration techniques. For example, common methods include static zero-bias compensation, earth rotation compensation, standard displacement table calibration, etc., and these methods are usually based on the laboratory environment and carried out under ideal static conditions. However, in actual application scenarios, INS often operates in a dynamic environment. For example, there are violent vibrations during the flight of an unmanned aerial vehicle, and complex road conditions affect vehicle navigation. In this case, the effectiveness of static calibration methods is greatly reduced, and they cannot fully reflect the influence of the dynamic environment on sensor errors. Therefore, how to achieve joint calibration of the grating sensor and INS under dynamic conditions has become the current research frontier direction. Summary of the Invention

[0005] To solve the above problems, the present invention provides a calibration method and device for an inertial navigation system based on a grating sensor.

[0006] To achieve the above object, the technical solutions adopted by the present invention are as follows:

[0007] On the one hand, the present invention discloses a calibration method for an inertial navigation system based on a grating sensor, including:

[0008] Step 1: Dynamically adjust the sampling frequencies and data acquisition strategies of the grating sensor and the IMU through the environmental perception adaptive adjustment mechanism, the real-time sensor health assessment model, and the real-time motion state classifier.

[0009] Step 2: Based on the real-time data interaction between the grating sensor and the IMU, design an interactive dynamic error allocation algorithm, dynamically adjust the sensor error correction weights according to the motion state and environmental changes, optimize the error compensation effect, and provide an error correction strategy for calibration.

[0010] Step 3: Use the adaptive partial differential equation algorithm to dynamically adjust the error compensation strategy by real-time correcting the partial differential equation and the adaptive correction coefficient of the error model.

[0011] Step 4: Introduce the virtual feedback correction loop mechanism and the deep reinforcement learning strategy, use the virtual environment model and the real-time feedback signal to optimize the calibration behavior, dynamically adjust the sensor weights, and achieve adaptive calibration.

[0012] Step 5: Adopt an incremental self-learning calibration mechanism, perform adaptive incremental optimization based on the error accumulation monitoring and historical data during operation, and dynamically adjust the calibration parameters.

[0013] Furthermore: Step 1 includes:

[0014] Real-time monitor environmental changes through the environmental perception module, and dynamically adjust the sampling frequencies of the grating sensor and the IMU according to the intensity of environmental changes.

[0015] Quantify the sensor health status based on the offset, drift, and noise level indicators of the sensor; if the health is below the threshold, mark the low-quality data through the health alert mechanism, and adjust the sampling frequency or trigger recalibration.

[0016] Use machine learning methods to real-time identify the current motion state, and adjust the sampling frequency and calibration strategy according to different motion states.

[0017] Furthermore: Step 2 includes:

[0018] Real-time judge the current motion state to decide how to adjust the sensor error correction weights.

[0019] Dynamically adjust the error correction ratios of the grating sensor and the IMU according to the real-time motion state. In the rotational motion stage, rely on the IMU data for correction, and in the linear motion stage, rely on the grating sensor data for correction.

[0020] Based on real-time sensor data and historical error data, the dynamic error allocation model algorithm monitors the errors of grating sensors and IMUs, and optimizes them using an adaptive weight adjustment model; when a large sensor error is detected, the correction weight of that sensor is increased, and the correction ratio of the other sensor is correspondingly decreased.

[0021] Further: Step 3 includes:

[0022] Partial differential equations are introduced to describe the errors of grating sensors and IMUs in time and space, so as to dynamically reflect the change of sensor errors over time, and at the same time consider the interaction between sensors;

[0023] An adaptive correction factor is introduced to adjust the error correction intensity in real time. When the acceleration or angular velocity changes suddenly, the correction ratio of the IMU is preferentially adjusted; when the environment changes little, the correction focus turns to the grating sensor;

[0024] The finite difference method is used to discretize and solve the partial differential equations to realize the dynamic adjustment of the error correction coefficient; by setting the time step, the error correction model is updated within each time step to track and correct the sensor errors in real time; at the same time, a dynamic error correction rate adjustment mechanism is introduced to adjust the correction speed according to the overall error feedback; when the error reaches the set threshold, the correction rate is accelerated; when the error is small, the correction rate is moderately slowed down to avoid error oscillations caused by overcorrection.

[0025] Further: Step 4 includes:

[0026] A virtual environment model is constructed to simulate the influence of various factors in the real environment on the sensor output; by comparing the actual output of the sensor with the expected output of the virtual model, an error feedback signal is generated; the feedback signal is transmitted to the artificial intelligence algorithm for simulating and optimizing the next calibration behavior;

[0027] The deep reinforcement learning algorithm is used to adaptively adjust the calibration strategy according to real-time errors and environmental feedback; the state space includes the current output of the sensor, error feedback, sensor health, acceleration, and angular velocity; the action space includes the intensity, direction, and weight allocation of sensor error correction; the reward function is defined as the reduction of the error, and the calibration process is optimized by maximizing the long-term reward; deep reinforcement learning determines the correction action through the policy network and evaluates the expected effect through the value network, so as to achieve dynamic optimization.

[0028] Further: Step 5 includes:

[0029] Periodically monitor the change of system error through an incremental calculation method based on a time window; set a dynamic time window, accumulate and statistically analyze the errors within each calibration period. If the accumulated error reaches a preset threshold, trigger the self-learning update mechanism and initiate the incremental adjustment of calibration parameters.

[0030] Based on historical data and real-time data, initiate an incremental optimization process. By analyzing the changing trend of historical errors, use a trend prediction algorithm to predict future error changes and dynamically adjust calibration parameters.

[0031] On the other hand, the present invention discloses an inertial navigation system calibration device based on a grating sensor, including:

[0032] Environmental perception and raw data acquisition module: Dynamically adjust the sampling frequencies and data acquisition strategies of the grating sensor and the IMU through an environmental perception adaptive adjustment mechanism, a real-time sensor health assessment model, and a real-time motion state classifier.

[0033] Interactive dynamic error allocation module: Based on the real-time data interaction between the grating sensor and the IMU, design an interactive dynamic error allocation algorithm, dynamically adjust the sensor error correction weights according to the motion state and environmental changes, optimize the error compensation effect, and provide an error correction strategy for calibration.

[0034] Dynamic error correction module: Utilize the adaptive partial differential equation algorithm to dynamically adjust the error compensation strategy by real-time correcting the partial differential equations of the error model and the adaptive correction coefficients.

[0035] Adaptive calibration module: Introduce a virtual feedback correction loop mechanism and a deep reinforcement learning strategy, optimize the calibration behavior using a virtual environment model and real-time feedback signals, dynamically adjust the sensor weights, and achieve adaptive calibration.

[0036] Calibration parameter optimization module: Adopt an incremental self-learning calibration mechanism, perform adaptive incremental optimization based on the error accumulation monitoring during operation and historical data, and dynamically adjust calibration parameters.

[0037] Furthermore: The environmental perception and raw data acquisition module includes:

[0038] Perceived environmental adaptive adjustment: Real-time monitor environmental changes through an environmental perception module, and dynamically adjust the sampling frequencies of the grating sensor and the IMU according to the intensity of environmental changes.

[0039] Real-time sensor health assessment module: Quantify the sensor health status based on the offset, drift, and noise level indicators of the sensor; if the health level is lower than the threshold, mark low-quality data through a health alert mechanism, and adjust the sampling frequency or trigger recalibration.

[0040] Real-time motion state classification module: Using machine learning methods, it can identify the current motion state in real time and adjust the sampling frequency and calibration strategy according to different motion states;

[0041] The interactive dynamic error allocation module includes:

[0042] Motion state recognition and sensor interaction analysis module: It can judge the current motion state in real time and decide how to adjust the error correction weight of the sensor;

[0043] Error allocation model construction module: According to the real-time motion state, it can dynamically adjust the error correction ratio of the grating sensor and the IMU. In the rotational motion stage, it depends on the IMU data for correction. In the linear motion stage, it depends on the grating sensor data for correction;

[0044] Real-time adjustment mechanism module of dynamic error correction weight: Based on real-time sensor data and historical error data, the dynamic error allocation model algorithm monitors the errors of the grating sensor and the IMU, and optimizes them using an adaptive weight adjustment model; when a large sensor error is detected, increase the correction weight of this sensor and correspondingly reduce the correction ratio of the other sensor.

[0045] Furthermore: The dynamic error correction module includes:

[0046] Partial differential equation construction module of error model: Introduce partial differential equations to describe the errors of the grating sensor and the IMU in time and space, so as to dynamically reflect the change of sensor errors over time, and at the same time consider the interaction between sensors;

[0047] Adaptive adjustment module of error correction coefficient: Introduce an adaptive correction factor to adjust the error correction intensity in real time. When the acceleration or angular velocity changes suddenly, give priority to adjusting the correction ratio of the IMU; when the environment changes little, the correction focus turns to the grating sensor;

[0048] Real-time solution of partial differential equations and system update module: Use the finite difference method to discretize and solve the partial differential equations to achieve dynamic adjustment of the error correction coefficient; by setting the time step, update the error correction model within each time step to track and correct sensor errors in real time; at the same time, introduce a dynamic error correction rate adjustment mechanism to adjust the correction speed according to the overall error feedback; when the error reaches the set threshold, the correction rate speeds up; when the error is small, the correction rate slows down moderately to avoid error oscillation caused by overcorrection.

[0049] Furthermore: The adaptive calibration module includes:

[0050] Virtual feedback correction loop mechanism module: Construct a virtual environment model to simulate the influence of various factors in the real environment on the sensor output; generate an error feedback signal by comparing the actual output of the sensor with the expected output of the virtual model; transmit the feedback signal to the artificial intelligence algorithm for simulating and optimizing the next calibration behavior;

[0051] Deep reinforcement learning strategy module: Use the deep reinforcement learning algorithm to adaptively adjust the calibration strategy according to the real-time error and environmental feedback; the state space includes the current output of the sensor, error feedback, sensor health, acceleration, and angular velocity; the action space includes the intensity, direction, and weight allocation of sensor error correction; the reward function is defined as the reduction of the error, and the calibration process is optimized by maximizing the long-term reward; deep reinforcement learning determines the correction action through the policy network and evaluates the expected effect through the value network, so as to achieve dynamic optimization;

[0052] The calibration parameter optimization module includes:

[0053] Error accumulation monitoring and self-learning trigger module: Periodically monitor the change of the system error through an incremental calculation method based on a time window; set a dynamic time window, accumulate and statistically analyze the error within each calibration cycle, and if the accumulated error reaches the preset threshold, trigger the self-learning update mechanism and start the incremental adjustment of the calibration parameters;

[0054] Incremental optimization strategy module: Based on historical data and real-time data, start the incremental optimization process, predict the future error change by analyzing the change trend of historical errors using a trend prediction algorithm, and dynamically adjust the calibration parameters.

[0055] Compared with the prior art, the technical progress achieved by the present invention is as follows:

[0056] Traditional INS calibration methods are usually based on static or quasi-static environments and are difficult to cope with the error changes in dynamic environments. This method adopts a perception environment adaptive adjustment mechanism, which can adjust the data acquisition frequency in real time and analyze the current motion state in combination with a real-time motion state classifier to achieve adaptive adaptation to dynamic environments. Traditional methods mostly use fixed weights to correct the errors of the data of IMU and grating sensors and are difficult to adapt to the complex changes in dynamic environments. This method introduces an interactive dynamic error allocation algorithm, which can dynamically allocate the error correction ratio according to the real-time data interaction relationship between the grating sensor and IMU. For example, during rotation, the data of the grating sensor is greatly affected by the moment of inertia, while in linear motion, the error of the IMU is easy to amplify. This algorithm can intelligently adjust the correction weight to achieve the optimal error compensation between multiple sensors.

[0057] Traditional error compensation methods mainly rely on discrete methods such as global optimization or Kalman filtering, and it is difficult to handle continuous error changes in complex dynamic environments. This method proposes an adaptive partial differential equation algorithm, which can dynamically adjust the error correction model in the continuous time domain by solving nonlinear partial differential equations. The adaptive partial differential equation algorithm combines the real-time motion state of the system and automatically adjusts the correction coefficient, making the error correction more accurate, especially suitable for high-speed changing environments, such as scenarios where drones quickly maneuver or vehicles make sharp turns. Traditional calibration methods are often carried out offline and are difficult to adapt to real-time error changes. This method proposes a virtual feedback correction loop, which constructs a virtual environment model, compares sensor errors with ideal outputs, and uses a deep reinforcement learning strategy for error trend prediction and dynamic optimization. The virtual feedback correction loop can continuously simulate future error changes and optimize the compensation strategy in advance, so as to achieve adaptive real-time calibration and improve the accuracy stability of INS during long-term operation.

[0058] In summary, the present invention extends from static calibration to dynamic calibration, from fixed error correction to interactive error allocation, upgrades from discrete error compensation to partial differential equation dynamic correction, and through the virtual feedback correction loop and self-learning mechanism, realizes the continuous optimization of the system, breaks through the limitations of static calibration, and can achieve high-precision error compensation in complex dynamic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification, and are used to explain the present invention together with the embodiments of the present invention, and do not constitute a limitation to the present invention.

[0060] In the drawings:

[0061] Figure 1 is a flowchart of Embodiment 1 of the present invention;

[0062] Figure 2 is a system structure diagram of Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present invention will be described below with reference to the drawings.

[0064] Embodiment 1

[0065] As Figure 1 shown, the present invention discloses a calibration method for an inertial navigation system based on a grating sensor, including: Step 1: Dynamically adjust the sampling frequencies and data acquisition strategies of the grating sensor and the IMU through a perception environment adaptive adjustment mechanism, a real-time sensor health assessment model, and a real-time motion state classifier;

[0066] Step 2: Based on the real-time data interaction between the grating sensor and the IMU, design an interactive dynamic error allocation algorithm to dynamically adjust the sensor error correction weights according to the motion state and environmental changes, optimize the error compensation effect, and provide an error correction strategy for calibration;

[0067] Step 3: Utilize the adaptive partial differential equation algorithm to dynamically adjust the error compensation strategy by real-time correcting the partial differential equation of the error model and the adaptive correction coefficient;

[0068] Step 4: Introduce a virtual feedback correction loop mechanism and a deep reinforcement learning strategy to optimize the calibration behavior using the virtual environment model and real-time feedback signals, dynamically adjust the sensor weights, and achieve adaptive calibration;

[0069] Step 5: Adopt an incremental self-learning calibration mechanism to perform adaptive incremental optimization based on the error accumulation monitoring during operation and historical data, and dynamically adjust the calibration parameters.

[0070] Specifically, the goal of Step 1 is to obtain the original sensor data in the dynamic environment, provide the necessary input data for the subsequent calibration process, especially to effectively collect the data of the grating sensor and the inertial measurement unit (IMU) in the dynamic environment, and adjust the data acquisition strategy in real time through an intelligent mechanism to provide the most accurate and reliable original data for the subsequent error modeling and calibration process.

[0071] 1.1 Perception environment adaptive adjustment mechanism:

[0072] The core of the perception environment adaptive adjustment mechanism is to dynamically adjust the sampling frequencies of the grating sensor and the IMU according to the instantaneous changes in the environment (such as vibration, acceleration, temperature changes, etc.). In this way, the system can adaptively adjust the working parameters of the sensor under different environmental conditions to reduce redundant data when the environment is stable, and at the same time increase the sampling frequency when the environment changes violently to ensure the accuracy of the data.

[0073] Specific implementation:

[0074] 1.1.1 Environment monitoring and sampling adjustment strategy:

[0075] When the environmental conditions change (such as external vibration, acceleration, temperature changes, etc.), the perception environment adaptive adjustment mechanism detects the current environmental changes through the built-in environmental perception module of the sensor (such as temperature and humidity sensors, accelerometers, etc.).

[0076] When the environmental vibration intensifies or the acceleration changes significantly, the system will increase the sampling frequency to capture the rapid dynamic characteristics in the environmental changes. For example, in a strong vibration environment, the system may increase the sampling frequency of the grating sensor to 10 kHz and the sampling frequency of the IMU to 1 kHz; while in stable and smooth motion, the sampling frequency is reduced to 1 kHz and 100 Hz.

[0077] 1.2.2 Adaptive frequency adjustment:

[0078] According to the environmental changes and the changes in the sensor output, the environmental perception adaptive adjustment mechanism uses dynamic threshold judgment to automatically adjust the sampling frequency and data update frequency of the sensor. The frequency adjustment strategy comprehensively considers information such as the current acceleration, vibration intensity, and noise of the sensor output. For example, when the standard deviation of the acceleration is greater than a certain threshold, the environmental perception adaptive adjustment mechanism will automatically trigger the high-frequency sampling mode.

[0079] Set the environmental parameters as , including factors such as acceleration, vibration intensity, temperature, etc. The change formula of the sensor sampling frequency is:

[0080]

[0081] Among them, is the reference sampling frequency, is the sensitivity coefficient, is a measure of the environmental change, and according to the environmental change, is automatically adjusted.

[0082] 1.1.3 Data redundancy control:

[0083] When the environment is relatively stable, the system will enter the "low-frequency sampling mode" to reduce redundant data and extend the service life of the sensor. In the low-frequency sampling mode, the storage and transmission load of the data is low, which helps to reduce the occupancy of the storage space.

[0084] For the grating sensor, the sampling frequency in the stable state may be reduced to 1 time per second, while the sampling frequency of the IMU is reduced to about 10 Hz.

[0085] 1.2 Real-time sensor health assessment model:

[0086] To ensure the credibility and accuracy of the data, the real-time sensor health assessment model will evaluate the operating status of the grating sensor and the IMU at every moment of data acquisition, and determine whether the sensor is working properly and effectively. This model will dynamically adjust the data acquisition strategy based on the real-time detected data characteristics to further improve the robustness of the system.

[0087] 1.2.1 Health Assessment Framework:

[0088] The sensor health assessment model relies on a set of metrics to quantify the health status of the sensor. These metrics include the offset, drift, noise level of the sensor, and the deviation of the sensor from the ideal value.

[0089] Assume the health of the grating sensor and the health of the IMU at time are defined by the following formulas respectively:

[0090]

[0091]

[0092] where, is the offset of the grating sensor, is the drift of the IMU, and are constants determined according to the sensor performance.

[0093] If the health assessment value is lower than a certain threshold, the system will mark the current data as low-quality data through the "health alert mechanism", and adjust the sampling frequency in a timely manner, or trigger a recalibration process.

[0094] 1.2.2 Sensor Health Status Monitoring:

[0095] The result of sensor health assessment is used to guide the change of sensor sampling frequency. When the health of the sensor is high, the system can select a lower sampling frequency to reduce the waste of computing resources. When the health of the sensor is low, the system will increase the sampling frequency or adopt a higher-precision calibration method.

[0096] 1.3 Real-time Motion State Classifier:

[0097] To effectively identify the current motion state (such as uniform linear motion, acceleration, rotation, etc.), this embodiment designs a "real-time motion state classifier" based on machine learning. The classifier analyzes the sensor data to identify the current motion state in real time, so as to adaptively adjust the sampling frequency and calibration process under different motion states.

[0098] 1.3.1 State Classification Model:

[0099] Use machine learning methods such as deep neural networks (DNN) or support vector machines (SVM) to train a classifier, and based on features such as the acceleration of the sensor, angular velocity, and displacement data of the grating sensor, predict the current motion state in real time. The model adopted in this embodiment will input the real-time data of the grating sensor and the IMU, and after preprocessing, input these features into the trained classifier to obtain the motion state prediction.

[0100] Assume the input features are (including acceleration, angular velocity, etc.), and the output of the classifier is the state category , which can be obtained through a deep learning network :

[0101]

[0102] where represents the th output node of the neural network, representing different motion states (such as uniform linear motion, acceleration, rotation, etc.).

[0103] 2. State adjustment logic:

[0104] After the classifier identifies the current state at each moment, it will output a motion state label, which will then trigger corresponding adaptive adjustment strategies. For example, during rotation, the system will pay more attention to the stability of the IMU; while during uniform motion, the accuracy of the grating sensor will be given priority.

[0105] The system will adjust the data acquisition strategy in a timely manner according to the change of the motion state. For example, it will increase the sampling frequency in the state of acceleration or large vibration, and decrease the frequency in the uniform state.

[0106] In step 1, by introducing a perception environment adaptive adjustment mechanism, a real-time sensor health assessment model, and a real-time motion state classifier, the system realizes a sensitive response to environmental changes and motion states. The perception environment adaptive adjustment mechanism ensures the data acquisition efficiency of the sensor in different dynamic environments. The health assessment model judges the reliability of the sensor in real time and adjusts the data acquisition strategy in a timely manner, while the real-time motion state classifier ensures that the system can make optimal data acquisition decisions in different motion states.

[0107] The specific goal of step 2 is to design a new dynamic error allocation model through the real-time interactive analysis of the data of the grating sensor and the inertial navigation system (INS), so that the system can dynamically adjust the error correction ratio of each sensor according to the motion state, environmental changes, and the interaction characteristics between sensors, optimize the error compensation effect, and improve the overall accuracy of the system.

[0108] 2.1 Interactive dynamic error allocation algorithm:

[0109] Traditional error correction methods usually rely on linear superposition models, simply adding the errors of each sensor and then performing correction. However, this method fails to consider the dynamic interaction between sensors and the changes in the motion environment. To address this limitation, this embodiment proposes an "Interactive Dynamic Error Allocation Algorithm". The core idea of this algorithm is: according to the real-time data interaction relationship between the grating sensor and the IMU, the environmental state, and different motion stages, dynamically adjust the proportion of error correction for each sensor. In different motion states, the impacts of the errors of the grating sensor and the IMU on the final navigation accuracy are not equal. Therefore, the error correction weight needs to be adaptively adjusted according to the motion state at each moment.

[0110] 2.1.1 Motion State Recognition and Sensor Interaction Analysis:

[0111] Based on the real-time motion state classifier introduced in Step 1, the dynamic error allocation model algorithm can judge the motion state of the system in real time (such as uniform linear motion, acceleration, rotation, etc.). In different motion states, the error performances of the grating sensor and the IMU are different. For example, in rotational motion, the error of the grating sensor may be amplified due to the moment of inertia effect, while in linear uniform motion, the acceleration error of the IMU usually becomes dominant. The dynamic error allocation model determines the current motion mode through the output information of the motion state classifier, and then decides how to adjust the error correction weight of the sensor.

[0112] 2.1.2 Construction of the Error Allocation Model:

[0113] In each motion state, the dynamic error allocation model algorithm will dynamically adjust the error correction ratio of the grating sensor and the IMU. For example, in the rotational motion stage, the system may give priority to relying on IMU data for correction, while in the linear motion stage, it relies more on grating sensor data for correction. The dynamic error allocation model distributes the correction weight of each sensor through the following mathematical model based on real-time sensor data such as acceleration, rotational angular velocity, and displacement.

[0114] 2.1.3 Error Correction Weight Allocation Formula:

[0115] Assume the errors of the grating sensor and the IMU are and , respectively. The total error correction of the system will be composed of the weighted corrections of the two. The dynamic allocation of the weight not only depends on the real-time data of the sensor, but also is related to the motion state and environmental feedback information (such as vibration, temperature, etc.). The specific formula is as follows:

[0116]

[0117] Wherein: is the error correction weight of the grating sensor, is the error correction weight of the IMU, , ensuring that the sum of the correction weights of the two sensors is 1.

[0118] Dynamic weight update rule:

[0119] The weights and will be dynamically updated according to the current motion state , acceleration , angular velocity and the health of the sensors and , and the specific update rules are as follows:

[0120]

[0121] Wherein, and are functions calculated based on factors such as the motion state, acceleration, and rotation information of the sensors, and are defined as follows:

[0122]

[0123]

[0124] These two functions measure the relative importance of the grating sensor and the IMU in the current environment and motion state. The health values and will dynamically affect the values of these functions, thereby adjusting the weights of each sensor. For example, during accelerating motion, the IMU may be more important; while during rotational motion, the influence of the grating sensor may be greater.

[0125] 2.1.4 Error collaborative correction:

[0126] Once the weights and are calculated, the system can perform weighted correction on the errors of the grating sensor and the IMU, thereby obtaining the optimal error correction value. By continuously iterating and updating the weights and correction values, the system can dynamically adjust the error allocation strategy at different motion stages and environmental conditions, and finally achieve error collaborative correction of multiple sensors.

[0127] 2.2 Real-time adjustment mechanism for dynamic error correction weights:

[0128] To better adapt to different motion states and environmental changes, the dynamic error allocation model algorithm not only depends on the initial weight allocation rule, but also self-adjusts according to real-time sensor data and system errors. As the system runs over time, the error relationship between sensors may change. Therefore, an adaptive weight adjustment mechanism is needed to optimize the error correction process.

[0129] 2.2.1 Dynamic weight adaptive adjustment:

[0130] Based on real-time sensor data, the dynamic error allocation model algorithm continuously monitors the errors of the grating sensor and the IMU, and uses historical data to establish an adaptive weight adjustment model. When a large sensor error is detected, the system automatically increases the correction weight of that sensor and correspondingly reduces the correction ratio of the other sensor. This adaptive mechanism updates the weights in real time through the following formula:

[0131]

[0132]

[0133] Where, is the adjustment coefficient, and are the error change amounts of the grating sensor and the IMU, representing the error correction amplitude of the current sensor relative to the previous moment. In this way, the system can adjust the weight allocation in real time and optimize the error correction.

[0134] By designing an interactive dynamic error allocation algorithm, this embodiment realizes a mechanism for dynamically adjusting the error correction weights between the grating sensor and the IMU, enabling the system to intelligently optimize the error correction strategy under different motion states and environmental conditions, achieving the best error collaborative correction effect. This algorithm not only considers the interaction between sensors, but also makes adaptive adjustments according to the real-time sensor health status and motion state, providing a reliable basis for subsequent precise calibration.

[0135] Specifically, the goal of step 3 is to ensure that the system can continuously maintain high accuracy in different motion environments and overcome the limitations of traditional global optimization or filtering methods. An adaptive partial differential equation algorithm is proposed, which can dynamically adjust the error correction strategy of the system according to the real-time changing environment and sensor errors, ensuring the real-time adaptability and accuracy of the system in high-dynamic environments.

[0136] 3.1 Adaptive partial differential equation algorithm:

[0137] Traditional error correction methods such as global optimization or filtering methods usually assume that environmental changes are static or gradual, so they are not suitable for highly dynamic application scenarios, especially under high-speed acceleration, rapid rotation, or sudden environmental changes. For this reason, this embodiment proposes an error correction model based on partial differential equations - the adaptive partial differential equation algorithm. This algorithm has a high degree of dynamic adaptability and can dynamically adjust the error compensation model through real-time correction rules to cope with rapidly changing motion states and environmental factors.

[0138] 3.1.1 Construction of the partial differential equation of the error model:

[0139] Based on the data of the grating sensor and the IMU, this embodiment assumes that the error of the system has dynamic changes in time and space. Therefore, partial differential equations are introduced to describe the time change of the error and the spatial interaction between sensors. Set the state variables of the sensor error as and , and the dynamic changes of these two errors are described by the following partial differential equations:

[0140]

[0141]

[0142] Among them, and respectively represent the errors of the grating sensor and the IMU, and are the acceleration and angular velocity, represents the influencing factors of the external environment (such as temperature, vibration, etc.).

[0143] The functions and are non-linear functions based on factors such as sensor error, motion state, and environmental changes, which describe the dynamic process of sensor error changing with time. This embodiment assumes that these functions have different dynamic behaviors under specific motion states, so they will be adaptively adjusted under different environmental conditions.

[0144] 3.1.2 Adaptive adjustment of the error correction coefficient:

[0145] In a dynamic environment, the adjustment of the error correction coefficient must respond to environmental changes in real time, especially under rapidly changing motion states such as acceleration, deceleration, and rotation. The adaptive partial differential equation algorithm adjusts the correction coefficient through adaptive correction rules to ensure that the error compensation of each sensor remains optimal under instantaneous environmental changes. For example, when the acceleration suddenly changes, the system should give priority to adjusting the correction ratio of the IMU; when the error of the grating sensor increases, its correction weight should be adjusted.

[0146] Specifically, this embodiment introduces an adaptive correction factor , which depends on factors such as acceleration, angular velocity, and sensor health, and adjusts the error correction intensity of each sensor in real time. This factor is calculated by the following formula:

[0147]

[0148] where and represent the amplitudes of acceleration and angular velocity respectively, is a constant representing the correction benchmark when the environment is static. This factor is used to adjust the intensity of error correction. For example:

[0149] If becomes larger, it indicates that there are large acceleration or rotation changes in the motion state, and the system should increase the proportion of IMU correction;

[0150] If becomes smaller, it means that the environmental change is small, and the system should shift the correction focus to the grating sensor.

[0151] Meanwhile, the correction coefficient also takes into account the health of the sensor ( and ), and automatically reduces the correction proportion of the sensor when the sensor fails or the data quality deteriorates.

[0152] 3.1.3 Real-time solution of partial differential equations and system update:

[0153] To correct errors in real time in a dynamic environment, the adaptive partial differential equation algorithm needs to quickly solve partial differential equations to update the sensor error correction model in real time. In this embodiment, the finite difference method in numerical solutions is used to discretize and solve the above partial differential equations to achieve dynamic adjustment of the error correction coefficient.

[0154] Set the time step , and this embodiment approximately solves the partial differential equation through the following discretized form:

[0155]

[0156]

[0157] By continuously iterating the above discretized equation, the adaptive partial differential equation algorithm can update the error correction model at each time step, thereby tracking and correcting the error between the grating sensor and the IMU in real time.

[0158] 3.2 Adaptive error correction rate:

[0159] To further improve the correction ability of the system in a high-dynamic environment, the adaptive partial differential equation algorithm also introduces a dynamic error correction rate adjustment mechanism. This mechanism adjusts the correction speed within each time step based on the overall error feedback of the system to cope with large fluctuations in system errors.

[0160] When the system error reaches the set threshold, the correction rate will automatically increase to quickly correct the system state; conversely, when the error is small, the correction rate will be appropriately slowed down to avoid error oscillations caused by overcorrection. The correction rate can be adjusted by the following formula:

[0161]

[0162] where is the adjustment factor, is the total error at the current moment. Through this mechanism, the adaptive partial differential equation algorithm can adjust the intensity and speed of correction according to real-time error feedback, ensuring that the system always maintains a high level of accuracy in different environments.

[0163] By introducing the adaptive partial differential equation algorithm, this embodiment constructs an error correction model with high dynamic adaptability. This model can perform dynamic correction of errors through non-linear partial differential equations based on real-time motion states, sensor data, and environmental changes, and combines an adaptive correction factor and a correction rate mechanism to adjust the error correction strategy in real time, thus effectively coping with challenges in various dynamic environments and ensuring that the system always maintains a high level of accuracy and stability.

[0164] Specifically, the goal of step 4 is to perform real-time calibration through a virtual model and optimize the accuracy of sensor data through step-by-step correction in combination with real-time environmental feedback. This process can not only simulate the behavior of sensors in a changing environment but also intelligently adjust the calibration behavior of the system through a deep reinforcement learning strategy to achieve efficient and dynamic adaptive calibration.

[0165] 4.1 Introduction of the virtual feedback correction loop mechanism:

[0166] The virtual feedback correction loop mechanism creates a virtual environment model, enabling the system to generate a feedback signal in real time based on the deviation between the actual output and the expected output of the sensor at each moment. This feedback signal is not directly fed back to the control system but is simulated and adjusted through a set of artificial intelligence algorithms for a more accurate calibration process. Specifically, the workflow of the virtual feedback correction loop mechanism can be divided into the following steps:

[0167] 4.1.1 Virtual environment modeling:

[0168] In this embodiment, a virtual environment model is first constructed based on the working characteristics of the sensor, environmental changes, and the dynamic error model. . This model can simulate the influence of various factors (such as vibration, acceleration, temperature change, etc.) in the real environment on the sensor output, and provide an ideal output for subsequent error correction. The error of the virtual environment model is defined by the difference between the actual output of the sensor and the expected output:

[0169]

[0170] This error is the core information in the virtual feedback correction loop, reflecting the deviation between the sensor and the ideal environment model.

[0171] 4.1.2 Generation and simulation of virtual feedback signals:

[0172] Based on the error , this embodiment uses a feedback mechanism to generate a virtual feedback signal . This signal is used to guide the next correction strategy. Different from the traditional error feedback method, this feedback signal is not directly used in the control system, but is transmitted to an artificial intelligence system to simulate how to optimize the next calibration behavior.

[0173] 4.2 Deep reinforcement learning strategy:

[0174] Deep reinforcement learning is the core technology to implement the virtual feedback correction loop mechanism. It can adaptively adjust the calibration strategy according to real-time errors and environmental feedback. Deep reinforcement learning simulates and predicts the future state of the environment, and gradually optimizes the calibration strategy according to the feedback signal.

[0175] 4.2.1 Definition of state space:

[0176] In deep reinforcement learning, the state space of the system includes multi-dimensional factors such as the current output of the sensor, environmental state, sensor health, error feedback information, etc. In this embodiment, the state space is defined as:

[0177]

[0178] where, is the current output of the sensor, is the error from the virtual environment model, and respectively represent the health of the grating sensor and the IMU, and is acceleration and angular velocity information, etc. This multi-dimensional state space can reflect the complex interaction relationship between the sensor and the environment.

[0179] 4.1.2 Action Space Definition:

[0180] Action Space includes the decision space for the correction strategy, that is, the intensity, correction direction, correction coefficient, etc. of the sensor error correction. The action space can be expressed as:

[0181]

[0182] where and are the corrections to the grating sensor and IMU errors respectively, and are the weight assignments of the grating sensor and IMU.

[0183] 4.1.3 Reward Function and Optimization Objective:

[0184] The goal of deep reinforcement learning is to minimize the error in the virtual environment by adjusting the actions of the system and optimize the calibration accuracy. In this process, the system takes an action based on the current state and feedbacks a reward according to the correction result. The reward function is defined as the reduction of the system error:

[0185]

[0186] The goal is to gradually optimize the calibration process by maximizing the long-term reward This process is achieved by the combined action of the policy network and value network of deep reinforcement learning. The policy network determines the correction actions to be taken, while the value network evaluates the expected effect of taking this action in the current state.

[0187] 4.1.4 Training and Adjustment:

[0188] The system uses an environment simulator and historical data to train the deep reinforcement learning model. In each training step, the virtual feedback signal is used as input to drive the policy network to learn, and the learning process is optimized through the reward function. As the training progresses, the model can automatically adapt to the dynamic changes in the environment and accurately adjust the calibration parameters of the sensor.

[0189] 4.3 Adaptive Adjustment of Sensor Weights:

[0190] During the virtual feedback correction loop process, the system can adjust the weight allocation of different sensors in real time according to the performance of the environment and sensors, ensuring that the overall calibration accuracy of the system is maximized under specific motion and environmental conditions. For example, in an environment of high-speed acceleration, the IMU may exhibit higher accuracy than the grating sensor, while in low-speed stable motion, the grating sensor may perform better. Therefore, through the virtual feedback correction loop, the system will automatically adjust the weights of the sensors according to the current state and feedback signals and , enabling the optimal performance of each sensor in a dynamic environment.

[0191] The weight adjustment formula is as follows:

[0192]

[0193]

[0194] where, and are dynamic weight adjustment factors related to the performance of each sensor, which are adjusted in real time based on sensor health and virtual error feedback.

[0195] By introducing the virtual feedback correction loop mechanism and the deep reinforcement learning strategy (deep reinforcement learning), this embodiment constructs a highly adaptive calibration process. This process enables the system to maintain precise calibration accuracy in various dynamic environments through real-time simulation and optimization of calibration behaviors. Through the modeling of the virtual environment, the generation of error feedback, and the optimization of deep learning, the virtual feedback correction loop mechanism can not only predict and correct sensor errors but also adjust the weight allocation of sensors according to real-time feedback, ensuring the accuracy of the entire system in complex environments.

[0196] Specifically, the goal of step 5 is to continuously optimize the calibration results using an incremental self-learning mechanism during the long-term operation of the system, ensuring that the errors in the actual use of the system are continuously corrected, maintaining high precision while adapting to changes in the environment and sensor performance.

[0197] 5.1 Incremental self-learning calibration mechanism:

[0198] The core goal of the incremental self-learning calibration mechanism is to continuously monitor the accumulation of system errors and perform adaptive incremental optimization based on historical data and real-time feedback. This process is not limited to the update of traditional error compensation models but combines the feedback during the actual use of the system to dynamically adjust the parameters of the calibration strategy and algorithm to ensure the stability and high precision of the system during long-term operation.

[0199] 5.1.1 Error accumulation monitoring and self-learning trigger:

[0200] During the operation of the system, the incremental self-learning calibration mechanism will periodically detect the change of the system error and analyze the error trend. For the error monitoring, in this embodiment, an incremental calculation method based on a time window is adopted, and a dynamic time window is set. The errors within each calibration period are accumulated and statistically analyzed:

[0201]

[0202] Among them, represents the total error of the system at time , is the error accumulation within this time window. If the accumulated error reaches the preset threshold , the self-learning update mechanism is triggered to start the incremental adjustment of the calibration parameters.

[0203] 5.1.2 Incremental optimization strategy:

[0204] Once it is detected that the error accumulation reaches the trigger condition, the incremental self-learning calibration mechanism will start the incremental optimization process based on historical data (including calibration error, environmental feedback, sensor health, etc.) and real-time data. Specifically, the algorithm of the incremental self-learning calibration mechanism adopts an incremental learning method to gradually adjust the parameters of the calibration algorithm. In this embodiment, the following methods are used for optimization:

[0205] Error trend analysis: The system will analyze the change trend of historical errors and use a trend prediction algorithm to predict future error changes. This analysis can be achieved through a regression model (such as the regression of historical data of acceleration error, angular velocity error, grating sensor error, etc.).

[0206] Dynamic adjustment of calibration parameters: According to the error trend and the current error accumulation situation, the incremental self-learning calibration mechanism will perform incremental adjustment on the calibration parameters in the error compensation model. The adjustment method is carried out through the gradient descent algorithm or an adaptive method based on Bayesian inference. In each incremental update period , the incremental self-learning calibration mechanism calculates the increment of the parameter:

[0207]

[0208] Among them, is the learning rate, represents the sensitivity to the current error, that is, the gradient of the calibration parameter. Through this incremental adjustment, the calibration process can gradually optimize the calibration result without disturbing the system operation.

[0209] 5.1.3 Incremental error correction model:

[0210] During the incremental optimization process, the incremental self-learning calibration mechanism does not simply compensate for errors, but deeply integrates information such as interaction with the dynamic environment and changes in sensor performance to adjust multiple parameters in the calibration model. Specifically, the incremental self-learning calibration mechanism combines the grating sensor and IMU error models in the previous steps and dynamically corrects each error model during incremental updates. Assuming that the system error is the weighted sum of multiple sensor errors, the incremental optimization process can be expressed by the following equation:

[0211]

[0212] where is the weight of each sensor , is the error of the sensor . The incremental self-learning calibration mechanism dynamically adjusts the contribution ratio of different sensors during calibration by incrementally updating the weight of each sensor.

[0213] 5.2 Self-learning update and algorithm parameter adjustment:

[0214] The incremental self-learning calibration mechanism system not only makes incremental corrections to the error compensation model, but also gradually adjusts the core parameters of the algorithm through self-learning. Traditional calibration methods usually rely on static model parameters, while the incremental self-learning calibration mechanism can adaptively adjust algorithm parameters based on the actual operation history of the system and long-term data accumulation through the self-learning mechanism, ensuring the stability and accuracy of the system during long-term use.

[0215] 5.2.1 Feedback-based self-learning update:

[0216] At the end of each calibration cycle, the incremental self-learning calibration mechanism optimizes the algorithm parameters during the calibration process according to the current error and compensation results. This process is updated through feedback based on historical data (such as dynamically adjusting the learning rate of calibration), enabling the system to continuously self-adjust based on long-term data.

[0217]

[0218] where is the initial learning rate, is the adjustment factor. As the error accumulates, the learning rate will gradually decrease, reflecting that the system enters a more stable learning state.

[0219] 5.2.2 Adaptive algorithm parameter update:

[0220] In addition to the calibration parameters in the error compensation model, the incremental self-learning calibration mechanism also adaptively adjusts other core parameters of the algorithm according to the historical data of system operation, such as the weight allocation of sensors, the speed of error correction, etc. The update of these parameters depends not only on error feedback, but also on the dynamic changes of the environment and the specific conditions of system operation. For example, the grating sensor may perform well during low-speed stable movement, while the IMU is more accurate during high-speed movement. The incremental self-learning calibration mechanism dynamically adjusts the weights of each sensor according to this real-time information to ensure the calibration accuracy of the system in different scenarios, by dynamically adjusting the weight function to achieve:

[0221]

[0222] Among them, is a dynamically adjusted factor related to the performance of each sensor, which is updated in real time based on the historical trend of errors and the current operating environment.

[0223] By introducing the incremental self-learning calibration mechanism, this embodiment designs a calibration optimization mechanism that can adapt to environmental changes and sensor performance evolution during long-term operation. The incremental self-learning calibration mechanism system ensures that the system can continuously correct and optimize the calibration error during long-term operation and always maintain high precision through steps such as periodic error accumulation monitoring, incremental optimization strategy, and real-time feedback self-learning.

[0224] Embodiment 2

[0225] As Figure 2 shown, the present invention discloses an inertial navigation system calibration device based on a grating sensor, including:

[0226] Environmental perception and raw data acquisition module: Dynamically adjust the sampling frequency and data acquisition strategy of the grating sensor and the IMU through the environmental perception adaptive adjustment mechanism, real-time sensor health assessment model, and real-time motion state classifier;

[0227] Interactive dynamic error allocation module: Based on the real-time data interaction between the grating sensor and the IMU, design an interactive dynamic error allocation algorithm, dynamically adjust the sensor error correction weight according to the motion state and environmental changes, optimize the error compensation effect, and provide an error correction strategy for calibration;

[0228] Dynamic error correction module: Use the adaptive partial differential equation algorithm to dynamically adjust the error compensation strategy by real-time correcting the partial differential equation of the error model and the adaptive correction coefficient;

[0229] Adaptive Calibration Module: Introduce a virtual feedback correction loop mechanism and a deep reinforcement learning strategy, utilize the virtual environment model and real-time feedback signals to optimize the calibration behavior, dynamically adjust the sensor weights, and achieve adaptive calibration;

[0230] Calibration Parameter Optimization Module: Adopt an incremental self-learning calibration mechanism, based on the error accumulation monitoring and historical data during operation, perform adaptive incremental optimization, and dynamically adjust the calibration parameters.

[0231] Specifically, the environmental perception and raw data acquisition module is divided into three sub-modules: the perception environment adaptive adjustment module, the real-time sensor health assessment module, and the real-time motion state classification module. First, the perception environment adaptive adjustment module uses the environmental perception module built into the sensor (such as accelerometers, temperature and humidity sensors, etc.) to monitor environmental changes (such as vibration, acceleration, temperature, etc.) in real time. According to the intensity of environmental changes, dynamically adjust the sampling frequencies of the grating sensor and the IMU. For example, when the vibration is strong or the acceleration changes greatly, the sampling frequency of the grating sensor is increased to 10 kHz, and the sampling frequency of the IMU is increased to 1 kHz; while when the environment is stable, the sampling frequency is reduced to reduce redundant data and extend the sensor life. The change in the sampling frequency is achieved through dynamic threshold judgment, taking into account factors such as acceleration, vibration intensity, and sensor output noise.

[0232] Secondly, the real-time sensor health assessment module quantifies the sensor health status based on indicators such as the offset, drift, and noise level of the sensor. If the health status is lower than the threshold, the system marks low-quality data through the "health alert mechanism" and adjusts the sampling frequency or triggers recalibration to ensure the credibility and accuracy of the data.

[0233] Finally, the real-time sensor health assessment module includes a real-time motion state classifier. The real-time motion state classifier uses machine learning methods such as deep neural networks (DNN) or support vector machines (SVM) to identify the current motion state (such as uniform linear motion, acceleration, rotation, etc.) in real time based on the acceleration, angular velocity, and displacement data of the sensor. According to different motion states, the system adjusts the sampling frequency and calibration strategy. For example, in rotational motion, the stability of the IMU is given priority, while in uniform motion, the accuracy of the grating sensor is given priority.

[0234] Specifically, the interactive dynamic error allocation module is divided into three sub-modules: the motion state recognition and sensor interaction analysis module, the error allocation model construction module, and the real-time adjustment mechanism module for dynamic error correction weights. First, in the motion state recognition and sensor interaction analysis module, based on the real-time motion state classifier introduced in the environmental perception and raw data acquisition module, the system can judge the current motion state in real time (such as uniform linear motion, acceleration, rotation, etc.). Under different motion states, the error performances of the grating sensor and the IMU are different. For example, in rotational motion, the error of the grating sensor may be amplified due to the inertia effect, while in uniform linear motion, the acceleration error of the IMU usually becomes dominant. Through the output information of the motion state classifier, the system decides how to adjust the error correction weights of the sensors.

[0235] Secondly, in the error allocation model construction module, the system dynamically adjusts the error correction ratios of the grating sensor and the IMU according to the real-time sensor data such as acceleration, rotational angular velocity, and displacement. The dynamic allocation of weights depends not only on the real-time data of the sensors but also on the motion state and environmental feedback information (such as vibration, temperature, etc.). For example, during the rotational motion stage, the system may give priority to relying on the IMU data for correction, while during the linear motion stage, it relies more on the grating sensor data for correction.

[0236] Finally, in the real-time adjustment mechanism module for dynamic error correction weights, the system optimizes the error correction process through an adaptive weight adjustment mechanism. Based on the real-time sensor data and the historical data of the system error, the dynamic error allocation model algorithm continuously monitors the errors of the grating sensor and the IMU and optimizes them using the adaptive weight adjustment model. When a large sensor error is detected, the system automatically increases the correction weight of that sensor and correspondingly reduces the correction ratio of the other sensor.

[0237] Specifically, the dynamic error correction module is divided into three sub-modules: the partial differential equation construction module for the error model, the adaptive adjustment module for the error correction coefficient, and the real-time solution of the partial differential equation and system update module. First, in the partial differential equation construction module for the error model, the system assumes that the errors of the grating sensor and the IMU have dynamic changes in time and space and introduces partial differential equations to describe the evolution process of these errors. The partial differential equations consider the effects of acceleration, angular velocity, and external environmental factors (such as temperature, vibration, etc.) on the errors.

[0238] Secondly, in the adaptive adjustment module for the error correction coefficient, the system introduces an adaptive correction factor to adjust the error correction intensity in real time according to factors such as acceleration, angular velocity, and sensor health. For example, when the acceleration or angular velocity changes suddenly, the system will give priority to adjusting the correction ratio of the IMU; while when the environment changes little, the focus of correction turns to the grating sensor.

[0239] Finally, in the real-time solution and system update module of the partial differential equation, the system uses the finite difference method to discretize and solve the partial differential equation to achieve the dynamic adjustment of the error correction coefficient. By setting the time step, the system updates the error correction model within each time step, tracking and correcting the sensor error in real time.

[0240] Specifically, the adaptive calibration module is divided into two sub-modules: the virtual feedback correction loop mechanism module and the deep reinforcement learning strategy module. First, in the virtual feedback correction loop mechanism module, the system constructs a virtual environment model to simulate the influence of various factors (such as vibration, acceleration, temperature change, etc.) in the real environment on the sensor output. By comparing the actual output of the sensor with the expected output of the virtual model, an error feedback signal is generated. This feedback signal is not directly used to control the system, but is transmitted to the artificial intelligence algorithm for simulating and optimizing the next calibration behavior.

[0241] Secondly, in the deep reinforcement learning strategy module, the system uses the deep reinforcement learning algorithm to adaptively adjust the calibration strategy according to the real-time error and environmental feedback. The state space includes multi-dimensional information such as the current output of the sensor, error feedback, sensor health, acceleration, and angular velocity. The action space includes the intensity, direction, and weight allocation of sensor error correction. The reward function is defined as the reduction of the system error, and the calibration process is optimized by maximizing the long-term reward. Deep reinforcement learning determines the correction action through the policy network and evaluates the expected effect through the value network, thus achieving dynamic optimization.

[0242] Specifically, the calibration parameter optimization module includes two sub-modules: the error accumulation monitoring and self-learning trigger module and the incremental optimization strategy module. First, in the error accumulation monitoring and self-learning trigger module, the system periodically monitors the change of the system error through an incremental calculation method based on a time window. A dynamic time window is set to accumulate and statistically analyze the error within each calibration cycle. If the error cumulative amount reaches the preset threshold, the self-learning update mechanism is triggered to start the incremental adjustment of the calibration parameters.

[0243] Secondly, in the incremental optimization strategy module, the system starts an incremental optimization process based on historical data (including calibration error, environmental feedback, sensor health, etc.) and real-time data. By analyzing the change trend of historical errors, a trend prediction algorithm (such as a regression model) is used to predict future error changes and dynamically adjust the calibration parameters.

[0244] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A calibration method for an inertial navigation system based on a grating sensor, characterized in that: include: Step 1: Dynamically adjust the sampling frequency and data acquisition strategy of the grating sensor and IMU through the adaptive adjustment mechanism of the perceived environment, the real-time sensor health assessment model and the real-time motion state classifier; Step 2: Based on the real-time data interaction between the grating sensor and the IMU, an interactive dynamic error allocation algorithm is designed to dynamically adjust the sensor error correction weight according to the motion state and environmental changes, optimize the error compensation effect, and provide an error correction strategy for calibration, including: Determine the current motion state in real time and decide how to adjust the sensor's error correction weight; According to the real-time motion state, the error correction ratio of the grating sensor and IMU is dynamically adjusted. In the rotational motion stage, the correction is based on the IMU data, and in the linear motion stage, the correction is based on the grating sensor data. Based on real-time sensor data and historical error data, the dynamic error allocation model algorithm monitors the errors of the grating sensor and IMU, and optimizes them using an adaptive weight adjustment model; when it is detected that the error of either the grating sensor or the IMU reaches the set threshold, the correction weight of the sensor is increased, and the correction weight of the other sensor is reduced accordingly; Step 3: Using the adaptive partial differential equation algorithm, the error compensation strategy is dynamically adjusted by real-time correction of the partial differential equation of the error model and the adaptive correction coefficient, including: Partial differential equations are introduced to describe the errors of grating sensors and IMU in time and space to dynamically reflect the changes of sensor errors over time while considering the interactions between sensors. Introduce an adaptive correction factor to adjust the error correction strength in real time. When the acceleration or angular velocity changes suddenly, the correction weight of the IMU is adjusted first. When the adaptive correction factor becomes smaller, the center of gravity is corrected to the grating sensor. The finite difference method is used to discretize and solve the partial differential equation to achieve dynamic adjustment of the error correction coefficient. By setting the time step, the error correction model is updated in each time step to track and correct the sensor error in real time. At the same time, a dynamic error correction rate adjustment mechanism is introduced to adjust the correction speed according to the overall error feedback. When the error reaches the set threshold, the correction rate is accelerated; when the error does not reach the set threshold, the correction rate is slowed down to avoid error oscillation caused by over-correction. Step 4: Introduce a virtual feedback correction loop mechanism and deep reinforcement learning strategy, use the virtual environment model and real-time feedback signals to optimize the calibration behavior, dynamically adjust the sensor weights, and achieve adaptive calibration; Step 5: Adopt an incremental self-learning calibration mechanism to perform adaptive incremental optimization and dynamically adjust calibration parameters based on error accumulation monitoring and historical data during operation.

2. The method for calibrating an inertial navigation system based on a grating sensor according to claim 1, characterized in that: The step 1 comprises: The environmental perception module monitors environmental changes in real time and dynamically adjusts the sampling frequency of the grating sensor and IMU according to the intensity of environmental changes; Quantify the health of the sensor based on the sensor's offset, drift, and noise level indicators; if the health is below the threshold, mark the low-quality data through the health alert mechanism, and adjust the sampling frequency or trigger recalibration; Use machine learning methods to identify the current motion state in real time, and adjust the sampling frequency and calibration strategy according to different motion states.

3. The method for calibrating an inertial navigation system based on a grating sensor according to claim 2, characterized in that: The step 4 comprises: Construct a virtual environment model to simulate the impact of various factors in the real environment on the sensor output; generate an error feedback signal by comparing the actual output of the sensor with the expected output of the virtual model; the feedback signal is passed to the artificial intelligence algorithm to simulate and optimize the next calibration behavior; The deep reinforcement learning algorithm is used to adaptively adjust the calibration strategy according to the real-time error and environmental feedback. The state space includes the current output of the sensor, error feedback, sensor health, acceleration and angular velocity. The action space includes the intensity, direction and weight distribution of the sensor error correction. The reward function is defined as the reduction in error, and the calibration process is optimized by maximizing the long-term reward. Deep reinforcement learning determines the correction action through the policy network, and the value network evaluates the expected effect, thereby achieving dynamic optimization.

4. The method for calibrating an inertial navigation system based on a grating sensor according to claim 3, characterized in that: The step 5 comprises: Through the incremental calculation method based on the time window, the changes of system errors are monitored periodically; a dynamic time window is set to accumulate and statistically analyze the errors in each calibration cycle. If the accumulated error reaches the preset threshold, the self-learning update mechanism is triggered to start the incremental adjustment of the calibration parameters; Based on historical data and real-time data, an incremental optimization process is started. By analyzing the changing trend of historical errors, a trend prediction algorithm is used to predict future error changes and dynamically adjust the calibration parameters.

5. A grating sensor-based inertial navigation system calibration device, characterized in that: include: Environmental perception and raw data acquisition module: Dynamically adjust the sampling frequency and data acquisition strategy of grating sensors and IMU through adaptive adjustment mechanism of perceived environment, real-time sensor health assessment model and real-time motion state classifier; Interactive dynamic error allocation module: Based on the real-time data interaction between the grating sensor and the IMU, an interactive dynamic error allocation algorithm is designed to dynamically adjust the sensor error correction weight according to the motion state and environmental changes, optimize the error compensation effect, and provide an error correction strategy for calibration, including: Motion state recognition and sensor interaction analysis module: determines the current motion state in real time and decides how to adjust the sensor's error correction weight; Building blocks of the error allocation model: dynamically adjust the error correction ratio of the grating sensor and IMU according to the real-time motion state. In the rotational motion stage, the error correction is based on the IMU data, and in the linear motion stage, the error correction is based on the grating sensor data. Real-time adjustment mechanism module of dynamic error correction weight: Based on real-time sensor data and historical error data, the dynamic error allocation model algorithm monitors the errors of the grating sensor and IMU, and optimizes them using the adaptive weight adjustment model; when it is detected that the error of any sensor in the grating sensor or IMU reaches the set threshold, the correction weight of the sensor is increased, and the correction weight of the other sensor is reduced accordingly; Dynamic error correction module: Using the adaptive partial differential equation algorithm, the error compensation strategy is dynamically adjusted by real-time correction of the partial differential equation of the error model and the adaptive correction coefficient, including: The dynamic error correction module comprises: Partial differential equation building blocks of the error model: Partial differential equations are introduced to describe the errors of the grating sensor and IMU in time and space to dynamically reflect the changes in sensor errors over time while considering the interactions between sensors; Adaptive adjustment module of error correction coefficient: introduce adaptive correction factor to adjust the error correction strength in real time. When the acceleration or angular velocity changes suddenly, the correction weight of IMU is adjusted first. When the adaptive correction factor becomes smaller, the center of gravity is corrected to the grating sensor. Real-time solution and system update module of partial differential equations: The finite difference method is used to discretize and solve the partial differential equations to achieve dynamic adjustment of the error correction coefficient; by setting the time step, the error correction model is updated in each time step to track and correct the sensor error in real time; at the same time, a dynamic error correction rate adjustment mechanism is introduced to adjust the correction speed according to the overall error feedback; when the error reaches the set threshold, the correction rate is accelerated; when the error does not reach the set threshold, the correction rate is slowed down to avoid error oscillation caused by over-correction; Adaptive calibration module: introduces a virtual feedback correction loop mechanism and deep reinforcement learning strategy, uses virtual environment models and real-time feedback signals to optimize calibration behavior, dynamically adjusts sensor weights, and achieves adaptive calibration; Calibration parameter optimization module: It adopts an incremental self-learning calibration mechanism, performs adaptive incremental optimization based on error accumulation monitoring and historical data during operation, and dynamically adjusts calibration parameters.

6. The grating sensor-based inertial navigation system calibration device according to claim 5, characterized in that: The environment perception and raw data acquisition module includes: Perception environment adaptive adjustment module: The environment perception module monitors environmental changes in real time and dynamically adjusts the sampling frequency of the grating sensor and IMU according to the intensity of environmental changes; Real-time sensor health assessment module: quantifies sensor health based on sensor offset, drift, and noise level indicators; if the health is below the threshold, the health alarm mechanism marks low-quality data and adjusts the sampling frequency or triggers recalibration; Real-time motion state classification module: Use machine learning methods to identify the current motion state in real time, and adjust the sampling frequency and calibration strategy according to different motion states.

7. The grating sensor-based inertial navigation system calibration device according to claim 6, characterized in that: The adaptive calibration module comprises: Virtual feedback correction loop mechanism module: builds a virtual environment model to simulate the impact of various factors in the real environment on the sensor output; generates an error feedback signal by comparing the actual output of the sensor with the expected output of the virtual model; the feedback signal is passed to the artificial intelligence algorithm to simulate and optimize the next calibration behavior; Deep reinforcement learning strategy module: uses deep reinforcement learning algorithm to adaptively adjust the calibration strategy according to real-time error and environmental feedback; the state space includes the current output of the sensor, error feedback, sensor health, acceleration and angular velocity; the action space includes the intensity, direction and weight distribution of sensor error correction; the reward function is defined as the amount of error reduction, and the calibration process is optimized by maximizing long-term rewards; deep reinforcement learning determines the correction action through the policy network, and the value network evaluates the expected effect, thereby achieving dynamic optimization; The calibration parameter optimization module includes: Error accumulation monitoring and self-learning triggering module: Through the incremental calculation method based on the time window, the system error changes are periodically monitored; a dynamic time window is set to accumulate and statistically analyze the errors in each calibration cycle. If the accumulated error reaches the preset threshold, the self-learning update mechanism is triggered to start the incremental adjustment of the calibration parameters; Incremental optimization strategy module: Based on historical data and real-time data, the incremental optimization process is started. By analyzing the changing trend of historical errors, the trend prediction algorithm is used to predict future error changes and dynamically adjust the calibration parameters.

Images