Inertial measurement unit (IMU) perception resource runtime performance evaluation method based on space-time prediction

Abstract

The invention discloses a spatio-temporal prediction-based performance evaluation method for IMU (Inertial Measurement Unit) sensing resources during operation. The method comprises the following steps: collecting IMU sensor data; the IMU sensor data is preprocessed; constructing a mixed space-time prediction model composed of a long short-term memory network and a convolutional auto-encoder; the IMU perception resource demand and the performance expression at the future moment are predicted; dynamically monitoring the resource use condition and the performance index of the system; according to a real-time performance evaluation result, optimizing an IMU perception resource allocation and use strategy by dynamically adjusting a resource scheduling scheme; an optimization algorithm is applied, efficient utilization of resources is ensured, and resource waste and performance bottleneck are avoided; and outputting a final evaluation result and an optimization suggestion for guiding subsequent operation of the system and IMU perception resource management. Through combination of space-time prediction and IMU perception resource scheduling, the operation efficiency of IMU perception resources is remarkably improved, and the method is suitable for application requirements in a complex dynamic environment.

Description

Technical Field

[0001] The present invention belongs to the field of sensor technology, and in particular relates to a runtime performance evaluation method of an IMU perception resource based on spatiotemporal prediction. Background Art

[0002] Based on the existing technical background, IMU (Inertial Measurement Unit) is widely used in fields such as navigation, robotics, and drones. IMU sensors measure the motion state and attitude of the device through accelerometers, gyroscopes, and magnetometers. However, with the complexity of application scenarios, existing technologies face many challenges in the performance evaluation of IMU perception resources during runtime.

[0003] First, traditional IMU data processing methods mainly focus on static data analysis and cannot effectively cope with complex and dynamic environmental changes. This results in low IMU perception resource utilization and performance evaluation accuracy in real-time applications. Traditional methods usually rely on fixed model parameters, lack flexible adjustment mechanisms, and cannot be dynamically updated according to actual scenarios, thus affecting the performance of the overall system.

[0004] Secondly, most existing spatiotemporal prediction technologies are based on a single data source or a simple linear model, which fails to fully utilize the potential information of multi-sensor data. This approach ignores the spatiotemporal correlation between sensor data, resulting in limited accuracy and reliability of prediction results. Especially in complex environments, data from a single sensor is susceptible to noise and outliers, which affects the effectiveness of the prediction model.

[0005] In addition, in terms of IMU sensing resource optimization and scheduling, traditional methods usually adopt static or predefined IMU sensing resource allocation strategies, which cannot respond to the actual fluctuations in IMU sensing resource demand in a timely manner. This static allocation method may lead to waste of IMU sensing resources or performance bottlenecks, reducing the overall efficiency of the system. The existing technology lacks accurate prediction and real-time adjustment mechanisms for IMU sensing resource requirements, and fails to effectively solve the problem of low IMU sensing resource utilization.

[0006] At the same time, the fusion and processing of multi-sensor data is also a major difficulty in existing technologies. Due to the diversity of sensor types and different data characteristics, how to extract effective information from multi-sensor data, reduce noise interference, and improve the accuracy of spatiotemporal prediction has become a major challenge. Existing methods usually simply average multi-sensor data and fail to fully utilize the complementarity between sensors.

[0007] The above defects will lead to inefficient use of IMU sensing resources in practical applications, which not only wastes system resources, but also may lead to insufficient system performance at critical moments. Therefore, an innovative method is urgently needed to improve the overall operating efficiency of IMU sensing resources by combining spatiotemporal prediction, real-time performance evaluation, resource optimization and multi-sensor data fusion to meet the application needs in complex dynamic environments. Summary of the invention

[0008] In order to overcome the shortcomings of the prior art, the present invention provides an IMU perception resource runtime performance evaluation method based on spatiotemporal prediction, including collecting IMU sensor data; preprocessing IMU sensor data; constructing a hybrid spatiotemporal prediction model composed of a long short-term memory network and a convolutional autoencoder; predicting the IMU perception resource requirements and performance at future moments; dynamically monitoring the resource usage and performance indicators of the system; optimizing the allocation and use strategy of IMU perception resources by dynamically adjusting the resource scheduling scheme according to the real-time performance evaluation results; applying optimization algorithms to ensure efficient use of resources, avoid resource waste and performance bottlenecks; outputting final evaluation results and optimization suggestions to guide the subsequent operation of the system and IMU perception resource management. The present invention significantly improves the operating efficiency of IMU perception resources by combining spatiotemporal prediction with IMU perception resource scheduling, and is suitable for application needs in complex dynamic environments.

[0009] The technical solution adopted by the present invention to solve the technical problem is as follows:

[0010] Step 1: Collect IMU sensor data, where the IMU sensor includes an accelerometer, a gyroscope, and a magnetometer;

[0011] Step 2: Clean, filter and remove noise from IMU sensor data;

[0012] Step 3: Based on the historical IMU sensor data and the real-time IMU sensor data, a hybrid spatiotemporal prediction model consisting of a long short-term memory network and a convolutional autoencoder is constructed, and the hybrid spatiotemporal prediction model is trained using a transfer learning algorithm based on an attention mechanism;

[0013] Step 4: Use the constructed hybrid spatiotemporal prediction model to predict the IMU perception resource requirements and performance at future times to obtain spatiotemporal prediction results;

[0014] Step 5: During operation, the efficiency of IMU sensing resources is evaluated in real time based on the spatiotemporal prediction results, and the resource usage and performance indicators of the system are dynamically monitored to obtain real-time efficiency evaluation results;

[0015] Step 6: According to the real-time performance evaluation results, optimize the allocation and use strategy of IMU sensing resources by dynamically adjusting the resource scheduling plan;

[0016] Step 7: Output the final evaluation results and optimization suggestions.

[0017] Furthermore, the step 1 is specifically as follows:

[0018] Step 1-1: Obtain IMU sensor data, including accelerometer, gyroscope and magnetometer data, which are collected at fixed time intervals;

[0019] Step 1-2: Collect accelerometer data A(t) and record the linear acceleration in the three-axis directions ; Represent the acceleration components on the X, Y, and Z axes respectively;

[0020] Step 1-3: Collect gyroscope data G(t) and record the angular velocity in the three-axis direction ; Represent the angular velocity components on the X, Y, and Z axes respectively;

[0021] Step 1-4: Collect magnetometer data M(t) and record the magnetic field strength in the three-axis direction , Represent the angular velocity components on the X, Y, and Z axes respectively.

[0022] Furthermore, the step 2 is specifically as follows:

[0023] Step 2-1: Clean the accelerometer data A(t), gyroscope data G(t), and magnetometer data M(t) obtained by the IMU sensor, and use an anomaly detection algorithm based on statistical analysis to remove outliers and missing data;

[0024] Step 2-2: Apply an adaptive Kalman filter to filter the cleaned accelerometer data A(t), gyroscope data G(t) and magnetometer data M(t), and dynamically adjust the filter gain in combination with real-time environmental variables;

[0025] Step 2-3: Perform multi-sensor data fusion on the filtered accelerometer data A(t), gyroscope data G(t) and magnetometer data M(t), and use the self-attention mechanism to enhance the signal features. The fused data is:

[0026] ;

[0027] ;

[0028] in, Both are weights based on the attention mechanism; represents the fused data; , , Represent the filtered accelerometer, gyroscope, and magnetometer data respectively; denote the query vector, key vector and value vector respectively, represents the attention score calculated by the self-attention mechanism, represents the exponential function;

[0029] Step 2-4: Fusion data For denoising, an adaptive denoising algorithm based on a deep convolutional neural network is applied to extract multi-scale features and obtain the final feature representation after convolution and activation operations:

[0030] ;

[0031] in, (.) is the convolution operation, is the convolutional layer weight; represents the final feature representation after convolution and activation operations, represents the total number of convolutional layers, Indicates layer of convolutional layers, represents the convolution operation, represents the activation function;

[0032] Step 2-5: Output IMU sensor data after data cleaning, filtering, data fusion and denoising , calculated by multi-scale feature enhancement method:

[0033] ;

[0034] in, and Respectively represent the multi-scale analysis function and feature extraction function; represents the number of dimensions for multi-scale analysis, Represents the mth dimension of multiscale analysis.

[0035] Furthermore, the step 3 is specifically as follows:

[0036] Step 3-1: Preprocessed IMU sensor data The input is fed into the enhanced long short-term memory network, and the temporal features and spatial correlation are extracted by combining the spatiotemporal gating unit. The output sequence feature H(t) is expressed as:

[0037] ;

[0038] in, , , represents the weight matrix, which is used for the current input, the hidden state at the previous moment, and the output of the spatiotemporal gating unit. represents the activation function, represents the bias vector; STGU(S(t)) represents the spatiotemporal gating unit, and the calculation formula is as follows:

[0039] ;

[0040] in, express The weight matrix of represents the bias vector, represents the hyperbolic tangent activation function;

[0041] The enhanced long short-term memory network refers to the introduction of a spatiotemporal gating unit STGU into the long short-term memory network;

[0042] Step 3-2: Input H(t) into the multi-scale convolutional autoencoder to capture features of different scales and generate a compressed representation C(t):

[0043] ;

[0044] in, Represents different convolution kernel sizes The convolution operation, Represents the pooling operation, which is used to reduce the feature dimension;

[0045] Step 3-3: Combine the dynamic attention mechanism to adjust the attention weight according to the historical context :

[0046] ;

[0047] in, is a dynamic adjustment function, which represents the calculation of the dot product of the query vector Q and the key vector K; R(t) represents the time-related bias, which captures the historical context information; Indicates the dimension size of K;

[0048] Step 3-4: Compress the representation Applying attention weights , generating spatiotemporal feature representation :

[0049] ;

[0050] in, Represents the balance parameter, which is used to fuse the original features and weighted features;

[0051] Step 3-5: Establish a hybrid spatiotemporal prediction model M and introduce a multi-objective optimization loss function:

[0052] ;

[0053] in, represents the multi-objective optimization loss function, represents the predicted value of the model output, Y(t) represents the actual observed value, , , represents the balance parameter of the loss function, corresponding to the error term, L2 regularization, and L1 regularization respectively; Represents the L2 and L1 norm regularization terms of the model weights respectively; represents the total number of time steps, represents the number of weights involved in L2 regularization, Indicates the number of weights involved in L1 regularization;

[0054] Step 3-6: Output the optimized hybrid spatiotemporal prediction model M.

[0055] Furthermore, the step 4 is specifically as follows:

[0056] Step 4-1: Representing spatiotemporal features Perform a fully connected layer transformation and output the potential feature representation:

[0057] ;

[0058] in, is the weight matrix, is the bias vector; represents the potential feature representation;

[0059] Step 4-2: Input the potential feature Z(t) into the prediction module, which is divided into IMU perception resource demand prediction and performance prediction :

[0060] ;

[0061] ;

[0062] in, and is the weight matrix of each module, and is bias; , They represent the amount of resources required at future times and the performance at future times, respectively. , They represent mapping functions respectively;

[0063] The prediction module is constructed by adding a fully connected layer on the basis of the hybrid spatiotemporal prediction model;

[0064] Step 4-3: Introduce the adaptive learning rate optimization algorithm to minimize the loss function:

[0065] ;

[0066] Among them, R(t) and P(t) are the actual IMU perception resource requirements and performance respectively; and is the weight parameter used to balance the prediction error;

[0067] Step 4-4: Optimize model parameters through iterative training to ensure the prediction accuracy of model M, and finally output the predicted IMU perception resource requirements and performance .

[0068] Furthermore, the step 5 is specifically as follows:

[0069] Step 5-1: During operation, collect current IMU perception resource usage data in real time through the sensor network and performance indicator data ;

[0070] Step 5-2: Combine predicted IMU sensing resource requirements and performance , construct the real-time monitoring matrix M(t):

[0071] ;

[0072] in, Indicates resource usage:

[0073] ;

[0074] in, is the resource fluctuation coefficient, is the fluctuation frequency, is the phase shift; represents the amount of resources actually used at time t, represents the maximum available resources at time t;

[0075] Represents the IMU sensing resource requirement error, plus the time derivative term to capture the rate of change:

[0076]

[0077] is the performance error, is the variance influence coefficient; Indicates actual performance value; Indicates the variance of performance;

[0078] Step 5-3: Adopt the adaptive weight allocation algorithm to analyze the real-time monitoring matrix. The weight adjustment formula is:

[0079] ;

[0080] in, and To adaptively adjust parameters, dynamically update based on historical data; is the weight adjustment coefficient; represents the dynamic weight, Represents the historical value of the real-time monitoring matrix;

[0081] Step 5-4: Generate performance evaluation report based on real-time monitoring matrix and weights :

[0082] ;

[0083] in, is the time impact coefficient, which is used to adjust the historical weight impact; For historical time points;

[0084] Step 5-5: According to the results of the performance evaluation report, trigger the intelligent optimization module to automatically adjust the system resource configuration;

[0085] Step 5-6: Output the performance evaluation report and the result of step 5-5 to the user interface.

[0086] Furthermore, the step 7 is specifically as follows:

[0087] Step 7-1: Based on the performance evaluation report Determine the IMU perception resource utilization and bottleneck indicators, and define the optimization objective function:

[0088] ;

[0089] in, , , To balance the coefficient, adjust the weights of different objectives;

[0090] Step 7-2: Apply the particle swarm optimization algorithm to initialize the particle swarm. Each particle represents an IMU sensing resource configuration scheme, and the particle position is expressed as:

[0091] ;

[0092] in, is the allocation of the jth IMU perception resource, expressed as:

[0093] ;

[0094] in, represents the predicted demand for the jth resource at time t, represents the predicted demand for the kth resource at time t;

[0095] Step 7-3: Calculate the position and velocity update of each particle:

[0096] ;

[0097] ;

[0098] in, is the inertia weight, which adjusts the influence of the particle's previous velocity. , is the learning factor, which affects the particle’s dependence on its own experience and group experience; , is a random number between [0,1] to increase the randomness of particle search; is the best historical position of the particle, is the global best position, is the derivative coefficient, combined with the gradient information of the optimization objective function; represents the velocity of particle i at time t+1, represents the velocity of particle i at time t, represents the position of particle i at time t, represents the objective function value corresponding to each particle, represents the position of particle i at time t+1;

[0099] Step 7-4: Calculate the optimization objective function value based on the updated particle position and update the particle's historical optimal position and the global optimal position :

[0100] ;

[0101] ;

[0102] Step 7-5: Determine whether the stop condition is met. If so, output the optimal IMU perception resource configuration plan. ; Otherwise, return to step 7-3 and continue iterating:

[0103] ;

[0104] in, is the convergence threshold to ensure that the optimization process reaches stability; It represents the objective function value corresponding to the global optimal position found during the system iteration at the previous time t-1;

[0105] Step 7-6: Apply the optimal IMU perception resource configuration solution , dynamically adjust the system IMU perception resource allocation to ensure the utilization and performance optimization of IMU perception resources:

[0106] ;

[0107] in, is the optimized j-th IMU perception resource allocation;

[0108] Step 7-7: Output the results of step 7-6 to the user interface, where the output includes optimal IMU perception resource configuration and error analysis.

[0109] The beneficial effects of the present invention are as follows:

[0110] (1) The present invention constructs a hybrid spatiotemporal prediction model, combines a long short-term memory network and a convolutional autoencoder, and makes full use of historical and real-time IMU sensor data to improve the accuracy of IMU perception resource requirements and performance prediction. Based on real-time performance evaluation of spatiotemporal prediction results, the system IMU perception resource usage and performance indicators are dynamically monitored, enabling the system to respond to environmental changes in a timely manner and improving overall adaptive capabilities.

[0111] (2) The present invention uses a particle swarm optimization algorithm combined with an adaptive learning rate to adjust the IMU perception resource configuration scheme in real time, thereby ensuring the efficient use of IMU perception resources, reducing the waste of IMU perception resources, and avoiding performance bottlenecks. It also uses an adaptive Kalman filter and a multi-sensor data fusion method based on an attention mechanism to reduce the error and uncertainty of a single sensor data and improve the reliability of spatiotemporal prediction.

[0112] (3) The present invention realizes intelligent adjustment of system resources by analyzing the real-time monitoring matrix and triggering the intelligent optimization module, significantly improving the overall efficiency of IMU perception resources, and providing a user interface for optimization results and adjustment suggestions, so that users can intuitively understand the improvement of system performance and IMU perception resource configuration, facilitating subsequent management and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0113] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0114] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0115] One purpose of the present invention is to propose a runtime performance evaluation method for IMU perception resources based on spatiotemporal prediction. The present invention significantly improves the operating efficiency of IMU perception resources by combining spatiotemporal prediction with IMU perception resource scheduling, and is suitable for application requirements in complex dynamic environments.

[0116] refer to Figure 1 , a runtime performance evaluation method of IMU perception resources based on spatiotemporal prediction is applied to a complex human behavior recognition scenario. In this scenario, multiple smart devices (such as wearable devices, smartphones, fitness equipment, etc.) are equipped with IMU sensors to monitor human posture, motion, position and other parameters in real time. Due to the diversity of motion and environmental interference, traditional methods have many shortcomings in real-time IMU perception resource scheduling and performance evaluation.

[0117] In a human behavior recognition task, multiple smart devices are deployed to collect user behavior data and transmit it to the central processing system. These smart devices are equipped with IMU sensors, including accelerometers, gyroscopes, and magnetometers. The recognition scene has the characteristics of complex and changing movements, which places high demands on the accuracy and real-time performance of the device's data collection. Under traditional methods, the lack of responsiveness to real-time movement changes can easily lead to IMU perception resource waste and performance bottlenecks.

[0118] In the method of the present invention, the intelligent device collects data in real time through the IMU sensor, which includes linear acceleration, angular velocity and magnetic field strength. The multi-sensor data is fused by using an adaptive Kalman filter and a self-attention mechanism to eliminate noise and outliers, thereby improving the accuracy and reliability of the data. Based on historical and real-time data, a hybrid spatiotemporal prediction model is constructed to make high-precision predictions on future IMU sensing resource requirements and performance.

[0119] During the task, the particle swarm optimization algorithm is used to adjust the IMU perception resource configuration of the smart device in real time to cope with the rapid changes in action. Based on the real-time performance evaluation results, the model dynamically monitors the IMU perception resource utilization and performance indicators to ensure that the smart device maintains stable performance in complex action scenarios.

[0120] In the behavior recognition task, the smart device can not only cope with complex action changes more effectively, but also maintain efficient resource utilization and stable performance during the task duration. Traditional methods are prone to problems such as improper IMU perception resource scheduling and performance bottlenecks when the action changes greatly. This method effectively avoids these problems through real-time prediction and optimization, and improves the overall performance of the behavior recognition task.

[0121] In general, the method of the present invention fully utilizes the perception capabilities of IMU sensors in complex dynamic environments. Through spatiotemporal prediction, performance evaluation and intelligent optimization, it achieves efficient utilization of IMU perception resources and significant improvement of system performance, providing a solid technical guarantee for human behavior recognition tasks.

[0122] In summary, the present invention significantly improves the operating efficiency of IMU perception resources by combining spatiotemporal prediction with IMU perception resource scheduling, and is suitable for application requirements in complex dynamic environments.

Claims

1. A runtime performance evaluation method for IMU perception resources based on spatiotemporal prediction, characterized in that: The steps include: Step 1: Collect IMU sensor data, where the IMU sensor includes an accelerometer, a gyroscope, and a magnetometer; Step 2: Clean, filter and remove noise from IMU sensor data; Step 3: Based on the historical IMU sensor data and the real-time IMU sensor data, a hybrid spatiotemporal prediction model consisting of a long short-term memory network and a convolutional autoencoder is constructed, and the hybrid spatiotemporal prediction model is trained using a transfer learning algorithm based on an attention mechanism; Step 4: Use the constructed hybrid spatiotemporal prediction model to predict the IMU perception resource requirements and performance at future times to obtain spatiotemporal prediction results; Step 5: During operation, the efficiency of IMU sensing resources is evaluated in real time based on the spatiotemporal prediction results, and the resource usage and performance indicators of the system are dynamically monitored to obtain real-time efficiency evaluation results; Step 6: According to the real-time performance evaluation results, optimize the allocation and use strategy of IMU sensing resources by dynamically adjusting the resource scheduling plan; Step 7: Output the final evaluation results and optimization suggestions.

2. According to claim 1, a method for evaluating the runtime performance of IMU sensing resources based on spatiotemporal prediction is characterized in that: The step 1 is specifically as follows: Step 1-1: Obtain IMU sensor data, including accelerometer, gyroscope and magnetometer data, which are collected at fixed time intervals; Step 1-2: Collect accelerometer data A(t) and record the linear acceleration in the three-axis direction ; Represent the acceleration components on the X, Y, and Z axes respectively; Step 1-3: Collect gyroscope data G(t) and record the angular velocity in the three-axis direction ; Represent the angular velocity components on the X, Y, and Z axes respectively; Step 1-4: Collect magnetometer data M(t) and record the magnetic field strength in the three-axis direction , Represent the angular velocity components on the X, Y, and Z axes respectively.

3. According to claim 2, a method for evaluating the runtime performance of IMU sensing resources based on spatiotemporal prediction is characterized in that: The step 2 is specifically as follows: Step 2-1: Clean the accelerometer data A(t), gyroscope data G(t), and magnetometer data M(t) obtained by the IMU sensor, and use an anomaly detection algorithm based on statistical analysis to remove outliers and missing data; Step 2-2: Apply an adaptive Kalman filter to filter the cleaned accelerometer data A(t), gyroscope data G(t) and magnetometer data M(t), and dynamically adjust the filter gain in combination with real-time environmental variables; Step 2-3: Perform multi-sensor data fusion on the filtered accelerometer data A(t), gyroscope data G(t) and magnetometer data M(t), and use the self-attention mechanism to enhance the signal features. The fused data is: ； ； in, Both are weights based on the attention mechanism; represents the fused data; , , Represent the filtered accelerometer, gyroscope, and magnetometer data respectively; denote the query vector, key vector and value vector respectively, represents the attention score calculated by the self-attention mechanism, represents the exponential function; Step 2-4: Fusion data For denoising, an adaptive denoising algorithm based on a deep convolutional neural network is applied to extract multi-scale features and obtain the final feature representation after convolution and activation operations: ； in, (.) is the convolution operation, is the convolutional layer weight; represents the final feature representation after convolution and activation operations, represents the total number of convolutional layers, Indicates layer of convolutional layers, represents the convolution operation, represents the activation function; Step 2-5: Output IMU sensor data after data cleaning, filtering, data fusion and denoising , calculated by multi-scale feature enhancement method: ； in, and Represent the multi-scale analysis function and feature extraction function respectively; represents the number of dimensions for multi-scale analysis, Represents the mth dimension of multiscale analysis.

4. The method for evaluating the runtime performance of IMU sensing resources based on spatiotemporal prediction according to claim 3 is characterized in that: The step 3 is as follows: Step 3-1: Preprocessed IMU sensor data The input is fed into the enhanced long short-term memory network, and the temporal features and spatial correlation are extracted by combining the spatiotemporal gating unit. The output sequence feature H(t) is expressed as: ； in, , , represents the weight matrix, which is used for the current input, the hidden state at the previous moment, and the output of the spatiotemporal gating unit. represents the activation function, represents the bias vector; STGU(S(t)) represents the spatiotemporal gating unit, and the calculation formula is as follows: ； in, express The weight matrix of represents the bias vector, represents the hyperbolic tangent activation function; The enhanced long short-term memory network refers to the introduction of a spatiotemporal gating unit STGU into the long short-term memory network; Step 3-2: Input H(t) into the multi-scale convolutional autoencoder to capture features of different scales and generate a compressed representation C(t): ； in, Represents different convolution kernel sizes The convolution operation, Represents the pooling operation, which is used to reduce the feature dimension; Step 3-3: Combine the dynamic attention mechanism to adjust the attention weight according to the historical context : ； in, is a dynamic adjustment function, which represents the calculation of the dot product of the query vector Q and the key vector K; R(t) represents the time-related bias, which captures the historical context information; Indicates the dimension size of K; Step 3-4: Compress the representation Applying attention weights , generating spatiotemporal feature representation : ； in, Represents the balance parameter, which is used to fuse the original features and weighted features; Step 3-5: Establish a hybrid spatiotemporal prediction model M and introduce a multi-objective optimization loss function: ； in, represents the multi-objective optimization loss function, represents the predicted value of the model output, Y(t) represents the actual observed value, , , represents the loss function balance parameter, corresponding to the error term, L2 regularization and L1 regularization respectively; Represents the L2 and L1 norm regularization terms of the model weights respectively; represents the total number of time steps, represents the number of weights involved in L2 regularization, Indicates the number of weights involved in L1 regularization; Step 3-6: Output the optimized hybrid spatiotemporal prediction model M.

5. The method for evaluating the runtime performance of IMU sensing resources based on spatiotemporal prediction according to claim 4 is characterized in that: The step 4 is specifically as follows: Step 4-1: Representing spatiotemporal features Perform a fully connected layer transformation and output the potential feature representation: ； in, is the weight matrix, is the bias vector; represents the potential feature representation; Step 4-2: Input the potential feature Z(t) into the prediction module, which is divided into IMU perception resource demand prediction and performance prediction : ； ； in, and is the weight matrix of each module, and is bias; , They represent the amount of resources required at future times and the performance at future times, respectively. , They represent mapping functions respectively; The prediction module is constructed by adding a fully connected layer on the basis of the hybrid spatiotemporal prediction model; Step 4-3: Introduce the adaptive learning rate optimization algorithm to minimize the loss function: ； Among them, R(t) and P(t) are the actual IMU perception resource requirements and performance respectively; and is the weight parameter used to balance the prediction error; Step 4-4: Optimize model parameters through iterative training to ensure the prediction accuracy of model M, and finally output the predicted IMU perception resource requirements and performance .

6. The method for evaluating the runtime performance of IMU sensing resources based on spatiotemporal prediction according to claim 5, characterized in that: The step 5 is specifically as follows: Step 5-1: During operation, collect current IMU perception resource usage data in real time through the sensor network and performance indicator data ; Step 5-2: Combine predicted IMU sensing resource requirements and performance , construct the real-time monitoring matrix M(t): ； in, Indicates resource usage: ； in, is the resource fluctuation coefficient, is the fluctuation frequency, is the phase shift; represents the amount of resources actually used at time t, represents the maximum available resources at time t; Represents the IMU sensing resource requirement error, plus the time derivative term to capture the rate of change: ； is the performance error, is the variance influence coefficient; Indicates actual performance value; Indicates the variance of performance; Step 5-3: Adopt the adaptive weight allocation algorithm to analyze the real-time monitoring matrix. The weight adjustment formula is: ； in, and To adaptively adjust parameters, dynamically update based on historical data; is the weight adjustment coefficient; represents the dynamic weight, Represents the historical value of the real-time monitoring matrix; Step 5-4: Generate performance evaluation report based on real-time monitoring matrix and weights : ； in, is the time impact coefficient, which is used to adjust the historical weight impact; For historical time points; Step 5-5: According to the results of the performance evaluation report, trigger the intelligent optimization module to automatically adjust the system resource configuration; Step 5-6: Output the performance evaluation report and the result of step 5-5 to the user interface.

7. The method for evaluating the runtime performance of IMU sensing resources based on spatiotemporal prediction according to claim 6, characterized in that: The step 7 is specifically as follows: Step 7-1: Based on the performance evaluation report Determine the IMU perception resource utilization and bottleneck indicators, and define the optimization objective function: ； in, , , To balance the coefficient, adjust the weights of different objectives; Step 7-2: Apply the particle swarm optimization algorithm to initialize the particle swarm. Each particle represents an IMU sensing resource configuration scheme, and the particle position is expressed as: ； in, is the allocation of the jth IMU perception resource, expressed as: ； in, represents the predicted demand for the jth resource at time t, represents the predicted demand for the kth resource at time t; Step 7-3: Calculate the position and velocity update of each particle: ； ； in, is the inertia weight, which adjusts the influence of the particle's previous velocity. , is the learning factor, which affects the particle’s dependence on its own experience and group experience; , is a random number between [0,1] to increase the randomness of particle search; is the best historical position of the particle, is the global best position, is the derivative coefficient, combined with the gradient information of the optimization objective function; represents the velocity of particle i at time t+1, represents the velocity of particle i at time t, represents the position of particle i at time t, represents the objective function value corresponding to each particle, represents the position of particle i at time t+1; Step 7-4: Calculate the optimization objective function value based on the updated particle position and update the particle's historical optimal position and the global optimal position : ； ； Step 7-5: Determine whether the stop condition is met. If so, output the optimal IMU perception resource configuration plan. ; Otherwise, return to step 7-3 and continue iterating: ； in, is the convergence threshold to ensure that the optimization process reaches stability; It represents the objective function value corresponding to the global optimal position found during the system iteration at the previous time t-1; Step 7-6: Apply the optimal IMU perception resource configuration solution , dynamically adjust the system IMU perception resource allocation to ensure the utilization and performance optimization of IMU perception resources: ； in, is the optimized j-th IMU perception resource allocation; Step 7-7: Output the results of step 7-6 to the user interface, where the output includes optimal IMU perception resource configuration and error analysis.

Images