A multi-sensor-based vehicle body posture online early warning method, system and medium

Abstract

This invention relates to a multi-sensor-based online vehicle attitude warning method, system, and medium. The method includes: S1. While the vehicle is traveling on the road, real-time motion state data of the left front side of the vehicle is acquired based on an IMU sensor on the left front side, real-time motion state data of the right front side of the vehicle is acquired based on an IMU sensor on the right front side, real-time motion state data of the left rear side of the vehicle is acquired based on an IMU sensor on the left rear side, and real-time motion state data of the right rear side of the vehicle is acquired based on an IMU sensor on the right rear side; S2. Based on the motion state data of the left front side, right front side, left rear side, and right rear side of the vehicle. This invention not only improves the monitoring range of unmanned container trucks in ports, enhancing safety and prevention capabilities, but also reduces the possibility of accidents, ensuring the safety of port operations.

Description

Technical Field

[0001] This invention relates to the technical field of autonomous vehicles, and in particular to a method, system, and medium for online early warning of vehicle posture based on multiple sensors. Background Technology

[0002] In recent years, with the country's vigorous promotion of smart port construction, more and more ports have begun to use unmanned trucks as a link between container yards and container ships for cargo handling. Unmanned trucks are often used in conjunction with intelligent monitoring platforms, allowing operators to remotely view vehicle location information such as lane, target yard bay number, and operation type. Due to uneven road surfaces causing slopes and uneven cargo distribution within containers, the vehicle's posture can tilt left and right, creating a risk of overturning. Therefore, in addition to the aforementioned location information, the vehicle's posture information detection is also crucial for the overall system safety. Vehicle posture detection refers to using sensors mounted on the unmanned truck to detect the vehicle's posture during operation, and the posture processing and calculation system outputs information such as the vehicle's tilt angle, pitch angle, heading angle, and acceleration, which is vital to the overall system safety. The data source for vehicle posture detection generally relies on IMU sensors installed at the rear wheel axle. Based on the assumption of a rigid connection between the IMU and the vehicle body, the posture changes of the IMU approximate the changes in the vehicle's posture. This method can only characterize the posture of a specific point on the vehicle. However, unmanned trucks are different. Their bodies are too long, and the container cargo they carry often weighs more than 40 tons. The suspension system of the vehicle body changes significantly. Relying on a single IMU installed at the center of the rear axle of the wheel cannot accurately represent the changes in the vehicle body's attitude. Therefore, this invention integrates data from multiple IMUs installed around the vehicle body to accurately sense the attitude changes of various parts of the vehicle body and predict risks in advance. It can promptly detect abnormal situations of unmanned trucks during operation, help operators better understand the current status of the vehicle, and reduce the risk of safety accidents.

[0003] In the prior art, Chinese patent (application number: 201780051103.2, publication number: CN109642783A) discloses a vehicle attitude detection system for detecting the attitude of a vehicle traveling on a road surface covered with magnetic markers. This system includes: a sensor unit comprising multiple magnetic sensors arranged in the vehicle width direction to measure the lateral offset relative to the magnetic markers; and a control unit that calculates the difference in lateral offset measured by the sensor units at two separate locations in the vehicle's longitudinal direction relative to any one of the magnetic markers. However, this solution requires a significant investment of time and money to modify the environment and lay magnetic markers, resulting in high operating costs and difficulty in rapid application in existing ports, thus exhibiting poor portability. Furthermore, the magnetic marker sensors require regular inspection and replacement, increasing on-site operating costs, making it unsuitable for unmanned truck attitude detection. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method, system and medium for online early warning of vehicle posture based on multiple sensors, which can not only improve the monitoring range of unmanned container trucks in ports, improve safety and prevention capabilities, but also reduce the possibility of accidents and ensure the safety of port operations.

[0005] To achieve the above and other related objectives, the present invention provides the following technical solution: an online vehicle posture warning method based on multiple sensors, the method comprising:

[0006] S1. When the vehicle is driving on the road, the motion status data of the left front side of the vehicle is obtained in real time based on the IMU sensor on the left front side of the vehicle, the motion status data of the right front side of the vehicle is obtained in real time based on the IMU sensor on the right front side of the vehicle, the motion status data of the left rear side of the vehicle is obtained in real time based on the IMU sensor on the left rear side of the vehicle, and the motion status data of the right rear side of the vehicle is obtained in real time based on the IMU sensor on the right rear side of the vehicle.

[0007] S2. Based on the motion state data information of the left front side, right front side, left rear side and right rear side of the vehicle body, a multi-source data fusion algorithm based on Pearson correlation coefficient is used to fuse the motion state of different sides of the vehicle body to obtain the fused motion state data information of the vehicle body.

[0008] S3. Based on the fused motion state data of the vehicle body, a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization is used to predict the attitude of the vehicle body, and the predicted attitude data of the vehicle body is obtained.

[0009] S4. Based on the predicted vehicle body posture data, construct the vehicle body posture evaluation function P, calculate the feasibility evaluation value of the vehicle body posture, and obtain the feasibility evaluation value data of the vehicle body posture.

[0010] Furthermore, the method also includes:

[0011] S5. Based on the feasibility assessment value data of the vehicle body posture, a preset threshold is set. If the feasibility assessment value of the vehicle body posture is less than the preset threshold, the vehicle will drive normally. If the feasibility assessment value of the vehicle body posture is greater than the preset threshold, the vehicle will roll over and a warning will be issued.

[0012] Furthermore, in step S2, the process of fusing the motion states of different sides of the vehicle body using a multi-source data fusion algorithm based on the Pearson correlation coefficient includes:

[0013] S21. Based on the motion state data of the left front side, right front side, left rear side, and right rear side of the vehicle body, establish the Pearson correlation function G of the vehicle body motion state.

[0014]

[0015] Where x1 represents the motion state data of the left front side of the vehicle body, x2 represents the motion state data of the right front side of the vehicle body, x3 represents the motion state data of the left rear side of the vehicle body, x4 represents the motion state data of the right rear side of the vehicle body, and α1, α2 and α3 are weighting factors of the vehicle body motion state. The correlation of the motion state of different sides of the vehicle body is characterized, and the correlation matrix of the motion state of different sides of the vehicle body is constructed to obtain the correlation matrix data of the motion state of different sides of the vehicle body.

[0016] S22. Based on the correlation matrix data of the motion states of different sides of the vehicle body, construct a fusion function H for the motion states of different sides of the vehicle body.

[0017]

[0018] Where y represents the correlation matrix data of the motion state of different sides of the vehicle body, and β1, β2 and β3 are the feature fusion factors of the motion state of different sides of the vehicle body.

[0019] S23. Based on the fusion function H of the motion states of different sides of the vehicle body, the motion states of different sides of the vehicle body are fused to obtain the motion state data information of the fused vehicle body.

[0020] Furthermore, the feature fusion factors β1, β2, and β3 for the motion states of different sides of the vehicle body are,

[0021]

[0022] Among them, x1 is the motion status data of the left front side of the vehicle body, x2 is the motion status data of the right front side of the vehicle body, x3 is the motion status data of the left rear side of the vehicle body, and x4 is the motion status data of the right rear side of the vehicle body.

[0023] Furthermore, the constraints for the weighting factors α1, α2, and α3 of the vehicle body motion state are as follows:

[0024]

[0025] Furthermore, in step S3, the prediction of the vehicle's attitude using a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization includes:

[0026] S31. Input the fused motion state data of the vehicle body into the gray neural network model for training and learning, initialize the network parameters of the gray neural network, and obtain the trained gray neural network model.

[0027] S32. Based on the trained gray neural network model, construct the network parameter matrix of the gray neural network model, initialize the sparrow population, determine the maximum number of iterations, and obtain the data information of the initialized sparrow population;

[0028] S33. Based on the data information of the initialized sparrow population, establish the fitness function Q for each individual in the population.

[0029]

[0030] Where z represents the data information of the sparrow population after initialization, and δ1 and δ2 are the fitness determinants of individual sparrows. The fitness values ​​of individual sparrows are calculated to obtain the fitness value data information of individual sparrows.

[0031] S34. Based on the fitness value data of the individual sparrows in the population, establish an objective optimization function W.

[0032]

[0033] Where r represents the fitness value data of individual sparrows in the population, and γ1, γ2 and γ3 are adaptive adjustment factors for target optimization. The network parameters of the gray neural network are optimized to obtain the optimized gray neural network model.

[0034] S35. Based on the optimized gray neural network model, input the fused motion state data of the vehicle body, predict the attitude of the vehicle body, and obtain the predicted attitude data of the vehicle body.

[0035] Furthermore, the constraint function f for the adaptive adjustment factors γ1, γ2, and γ3 of the target optimization is,

[0036]

[0037] The constraints on the fitness determinants of individuals in the sparrow population are as follows:

[0038]

[0039] The constraint function f takes values ​​in the range (3,5).

[0040] Furthermore, the attitude evaluation function P of the vehicle body is,

[0041]

[0042] Where g represents the predicted vehicle body posture data, and η1, η2, and η3 are the evaluation factors for vehicle body posture.

[0043] To achieve the above and other related objectives, the present invention also provides a computer-readable storage medium storing a computer program programmed or configured to perform any of the multi-sensor-based online vehicle attitude warning methods described in the present invention.

[0044] To achieve the above and other related objectives, the present invention also provides a computer-readable storage medium storing a computer program programmed or configured to perform any of the multi-sensor-based online vehicle attitude warning methods described in the present invention.

[0045] The present invention has the following positive effects:

[0046] 1. This invention uses a multi-source data fusion algorithm based on Pearson correlation coefficient to fuse the motion states of different sides of the vehicle body, and combines it with a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization to predict the vehicle body's attitude. This not only improves the monitoring range of unmanned container trucks in ports, enhances safety and prevention capabilities, but also reduces the possibility of accidents and ensures the safety of port operations.

[0047] 2. This invention constructs a vehicle body attitude evaluation function P to calculate the feasibility evaluation value of the vehicle body attitude, providing real-time feedback and predictive alarms. Operation supervisors can make timely adjustments, which not only maintains the stability of the vehicle during driving, but also allows the vehicle body attitude detection to diagnose faults in components such as the vehicle's suspension system and steering system, helping to identify and solve problems in a timely manner and improve the reliability and durability of the vehicle. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0049] Figure 2 This is a flowchart illustrating the multi-source data fusion algorithm based on Pearson correlation coefficient of the present invention.

[0050] Figure 3 This is a flowchart illustrating the regression prediction algorithm of the gray neural network based on multi-objective sparrow optimization of the present invention.

[0051] Figure 4 This is a schematic diagram of the system framework of the present invention. Detailed Implementation

[0052] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0053] Example 1: As Figure 1 As shown, a multi-sensor-based online vehicle attitude warning method is provided, the method comprising:

[0054] S1. When the vehicle is driving on the road, the motion status data of the left front side of the vehicle is obtained in real time based on the IMU sensor on the left front side of the vehicle, the motion status data of the right front side of the vehicle is obtained in real time based on the IMU sensor on the right front side of the vehicle, the motion status data of the left rear side of the vehicle is obtained in real time based on the IMU sensor on the left rear side of the vehicle, and the motion status data of the right rear side of the vehicle is obtained in real time based on the IMU sensor on the right rear side of the vehicle.

[0055] S2. Based on the motion state data information of the left front side, right front side, left rear side and right rear side of the vehicle body, a multi-source data fusion algorithm based on Pearson correlation coefficient is used to fuse the motion state of different sides of the vehicle body to obtain the fused motion state data information of the vehicle body.

[0056] S3. Based on the fused motion state data of the vehicle body, a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization is used to predict the attitude of the vehicle body, and the predicted attitude data of the vehicle body is obtained.

[0057] S4. Based on the predicted vehicle body posture data, construct the vehicle body posture evaluation function P, calculate the feasibility evaluation value of the vehicle body posture, and obtain the feasibility evaluation value data of the vehicle body posture.

[0058] In this embodiment, the method further includes:

[0059] S5. Based on the feasibility assessment value data of the vehicle body posture, a preset threshold is set. If the feasibility assessment value of the vehicle body posture is less than the preset threshold, the vehicle will drive normally. If the feasibility assessment value of the vehicle body posture is greater than the preset threshold, the vehicle will roll over and a warning will be issued.

[0060] In this embodiment, as Figure 2 As shown, in step S2, the process of fusing the motion states of different sides of the vehicle body using a multi-source data fusion algorithm based on the Pearson correlation coefficient includes:

[0061] S21. Based on the motion state data of the left front side, right front side, left rear side, and right rear side of the vehicle body, establish the Pearson correlation function G of the vehicle body motion state.

[0062]

[0063] Where x1 represents the motion state data of the left front side of the vehicle body, x2 represents the motion state data of the right front side of the vehicle body, x3 represents the motion state data of the left rear side of the vehicle body, x4 represents the motion state data of the right rear side of the vehicle body, and α1, α2 and α3 are weighting factors of the vehicle body motion state. The correlation of the motion state of different sides of the vehicle body is characterized, and the correlation matrix of the motion state of different sides of the vehicle body is constructed to obtain the correlation matrix data of the motion state of different sides of the vehicle body.

[0064] S22. Based on the correlation matrix data of the motion states of different sides of the vehicle body, construct a fusion function H for the motion states of different sides of the vehicle body.

[0065]

[0066] Where y represents the correlation matrix data of the motion state of different sides of the vehicle body, and β1, β2 and β3 are the feature fusion factors of the motion state of different sides of the vehicle body.

[0067] S23. Based on the fusion function H of the motion states of different sides of the vehicle body, the motion states of different sides of the vehicle body are fused to obtain the motion state data information of the fused vehicle body.

[0068] In this embodiment, the feature fusion factors β1, β2, and β3 of the motion states on different sides of the vehicle body are,

[0069]

[0070] Among them, x1 is the motion status data of the left front side of the vehicle body, x2 is the motion status data of the right front side of the vehicle body, x3 is the motion status data of the left rear side of the vehicle body, and x4 is the motion status data of the right rear side of the vehicle body.

[0071] In this embodiment, the constraints for the weighting factors α1, α2, and α3 of the vehicle body motion state are as follows:

[0072]

[0073] Example 2: Based on the multi-sensor-based online vehicle attitude warning method in Example 1, the present invention will be further explained and described below.

[0074] like Figure 1 As shown, a multi-sensor-based online vehicle attitude warning method is provided, the method comprising:

[0075] S1. When the vehicle is driving on the road, the motion status data of the left front side of the vehicle is obtained in real time based on the IMU sensor on the left front side of the vehicle, the motion status data of the right front side of the vehicle is obtained in real time based on the IMU sensor on the right front side of the vehicle, the motion status data of the left rear side of the vehicle is obtained in real time based on the IMU sensor on the left rear side of the vehicle, and the motion status data of the right rear side of the vehicle is obtained in real time based on the IMU sensor on the right rear side of the vehicle.

[0076] S2. Based on the motion state data information of the left front side, right front side, left rear side and right rear side of the vehicle body, a multi-source data fusion algorithm based on Pearson correlation coefficient is used to fuse the motion state of different sides of the vehicle body to obtain the fused motion state data information of the vehicle body.

[0077] S3. Based on the fused motion state data of the vehicle body, a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization is used to predict the attitude of the vehicle body, and the predicted attitude data of the vehicle body is obtained.

[0078] S4. Based on the predicted vehicle body posture data, construct the vehicle body posture evaluation function P, calculate the feasibility evaluation value of the vehicle body posture, and obtain the feasibility evaluation value data of the vehicle body posture.

[0079] In this embodiment, as Figure 3 As shown, in step S3, the prediction of the vehicle's attitude using a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization includes:

[0080] S31. Input the fused motion state data of the vehicle body into the gray neural network model for training and learning, initialize the network parameters of the gray neural network, and obtain the trained gray neural network model.

[0081] S32. Based on the trained gray neural network model, construct the network parameter matrix of the gray neural network model, initialize the sparrow population, determine the maximum number of iterations, and obtain the data information of the initialized sparrow population;

[0082] S33. Based on the data information of the initialized sparrow population, establish the fitness function Q for each individual in the population.

[0083]

[0084] Where z represents the data information of the sparrow population after initialization, and δ1 and δ2 are the fitness determinants of individual sparrows. The fitness values ​​of individual sparrows are calculated to obtain the fitness value data information of individual sparrows.

[0085] S34. Based on the fitness value data of the individual sparrows in the population, establish an objective optimization function W.

[0086]

[0087] Where r represents the fitness value data of individual sparrows in the population, and γ1, γ2 and γ3 are adaptive adjustment factors for target optimization. The network parameters of the gray neural network are optimized to obtain the optimized gray neural network model.

[0088] S35. Based on the optimized gray neural network model, input the fused motion state data of the vehicle body, predict the attitude of the vehicle body, and obtain the predicted attitude data of the vehicle body.

[0089] In this embodiment, the constraint function f for the adaptive adjustment factors γ1, γ2, and γ3 of the target optimization is,

[0090]

[0091] The constraints on the fitness determinants of individuals in the sparrow population are as follows:

[0092]

[0093] The constraint function f takes values ​​in the range (3,5).

[0094] In this embodiment, the attitude evaluation function P of the vehicle body is,

[0095]

[0096] Where g represents the predicted vehicle body posture data, and η1, η2, and η3 are the evaluation factors for vehicle body posture.

[0097] In this embodiment, the present invention provides a computer-readable storage medium storing a computer program programmed or configured to perform any of the multi-sensor-based online vehicle attitude warning methods described above.

[0098] In this embodiment, as Figure 4As shown, the vehicle's attitude in its unloaded state is recorded as the initial value at the moment of power-on. Information such as speed, acceleration, roll angle, and heading angle, acquired by IMUs installed at the four corners of the vehicle, is transmitted to the controller via wired transmission. The controller uses a data preprocessing algorithm to eliminate interference from abnormal data and complete time alignment. After receiving information from multiple IMU sensors, the fusion algorithm calculates the vehicle's pose and compensates for each incoming data by calculating the IMU zero bias value, obtaining the true angle change and predicting the vehicle's attitude change trend over the next 30 seconds. The current vehicle attitude information calculated by the fusion algorithm and the predicted vehicle change trend are compared with the unloaded state saved at the initial power-on moment. If the angle exceeds the set allowable change threshold, a warning message is issued. The calculated attitude information and alarm signal are uploaded to a remote monitoring platform via the vehicle's 5G wireless transmission facility. Platform personnel then promptly issue deceleration or stopping commands, which are transmitted to the vehicle via the 5G transmission communication module.

[0099] In this embodiment, the present invention provides a computer-readable storage medium storing a computer program programmed or configured to perform any of the multi-sensor-based online vehicle attitude warning methods described above.

[0100] Any references to memory, storage, database, or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0101] In summary, this invention not only enhances the monitoring range of unmanned container trucks in ports, improving safety and prevention capabilities, but also reduces the likelihood of accidents and ensures the safety of port operations.

[0102] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for online early warning of vehicle attitude based on multiple sensors, characterized in that, The method includes: S1. When the vehicle is driving on the road, the motion status data of the left front side of the vehicle is obtained in real time based on the IMU sensor on the left front side of the vehicle, the motion status data of the right front side of the vehicle is obtained in real time based on the IMU sensor on the right front side of the vehicle, the motion status data of the left rear side of the vehicle is obtained in real time based on the IMU sensor on the left rear side of the vehicle, and the motion status data of the right rear side of the vehicle is obtained in real time based on the IMU sensor on the right rear side of the vehicle. S2. Based on the motion state data information of the left front side, right front side, left rear side and right rear side of the vehicle body, a multi-source data fusion algorithm based on Pearson correlation coefficient is used to fuse the motion state of different sides of the vehicle body to obtain the fused motion state data information of the vehicle body. S3. Based on the fused motion state data of the vehicle body, a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization is used to predict the attitude of the vehicle body, and the predicted attitude data of the vehicle body is obtained. S4. Based on the predicted vehicle body posture data, construct the vehicle body posture evaluation function P, calculate the feasibility evaluation value of the vehicle body posture, and obtain the feasibility evaluation value data of the vehicle body posture. S5. Based on the feasibility assessment value data of the vehicle body posture, a preset threshold is set. If the feasibility assessment value of the vehicle body posture is less than the preset threshold, the vehicle will drive normally. If the feasibility assessment value of the vehicle body posture is greater than the preset threshold, the vehicle will roll over and a warning will be issued.

2. The online vehicle attitude warning method based on multiple sensors according to claim 1, characterized in that, In step S2, the process of fusing the motion states of different sides of the vehicle body using a multi-source data fusion algorithm based on the Pearson correlation coefficient includes: S21. Based on the motion state data of the left front side, right front side, left rear side, and right rear side of the vehicle body, establish the Pearson correlation function G of the vehicle body motion state. ， Where x1 represents the motion state data of the left front side of the vehicle body, x2 represents the motion state data of the right front side of the vehicle body, x3 represents the motion state data of the left rear side of the vehicle body, x4 represents the motion state data of the right rear side of the vehicle body, α1, α2 and α3 are weighting factors of the vehicle body motion state, the correlation of the motion state of different sides of the vehicle body is characterized, and the correlation matrix of the motion state of different sides of the vehicle body is constructed to obtain the correlation matrix data of the motion state of different sides of the vehicle body. S22. Based on the correlation matrix data of the motion states of different sides of the vehicle body, construct a fusion function H for the motion states of different sides of the vehicle body. ， Where y represents the correlation matrix data of the motion state of different sides of the vehicle body, and β1, β2 and β3 are the feature fusion factors of the motion state of different sides of the vehicle body. S23. Based on the fusion function H of the motion states of different sides of the vehicle body, the motion states of different sides of the vehicle body are fused to obtain the motion state data information of the fused vehicle body.

3. The online vehicle attitude warning method based on multiple sensors according to claim 2, characterized in that: The characteristic fusion factors β1, β2, and β3 of the motion states on different sides of the vehicle body are, ， ， ， Among them, x1 is the motion status data of the left front side of the vehicle body, x2 is the motion status data of the right front side of the vehicle body, x3 is the motion status data of the left rear side of the vehicle body, and x4 is the motion status data of the right rear side of the vehicle body.

4. The online vehicle attitude warning method based on multiple sensors according to claim 2, characterized in that: The constraints for the weighting factors α1, α2, and α3 of the vehicle body motion state are as follows: 。 5. The online vehicle attitude warning method based on multiple sensors according to claim 1, characterized in that, In step S3, the prediction of the vehicle's attitude using a regression prediction algorithm based on a gray neural network with multi-objective sparrow optimization includes: S31. Input the fused motion state data of the vehicle body into the gray neural network model for training and learning, initialize the network parameters of the gray neural network, and obtain the trained gray neural network model. S32. Based on the trained gray neural network model, construct the network parameter matrix of the gray neural network model, initialize the sparrow population, determine the maximum number of iterations, and obtain the data information of the initialized sparrow population; S33. Based on the data information of the initialized sparrow population, establish the fitness function Q for each individual in the population. ， Where z represents the data information of the sparrow population after initialization, and δ1 and δ2 are the fitness determinants of individual sparrows. The fitness values ​​of individual sparrows are calculated to obtain the fitness value data information of individual sparrows. S34. Based on the fitness value data of the individual sparrows in the population, establish an objective optimization function W. ， Where r represents the fitness value data of individual sparrows in the population, and γ1, γ2 and γ3 are adaptive adjustment factors for target optimization. The network parameters of the gray neural network are optimized to obtain the optimized gray neural network model. S35. Based on the optimized gray neural network model, input the fused motion state data of the vehicle body, predict the attitude of the vehicle body, and obtain the predicted attitude data of the vehicle body.

6. The online vehicle attitude warning method based on multiple sensors according to claim 5, characterized in that: The constraint function f for the adaptive adjustment factors γ1, γ2, and γ3 of the target optimization is, ， The constraints on the fitness determinants of individuals in the sparrow population are as follows: ， The constraint function f takes values ​​in the range (3,5).

7. The online vehicle attitude warning method based on multiple sensors according to claim 1, characterized in that: The attitude evaluation function P of the vehicle body is, ， Where g represents the predicted vehicle body posture data, and η1, η2, and η3 are the evaluation factors for vehicle body posture.

8. A multi-sensor-based online vehicle attitude warning system, comprising computer equipment, characterized in that, The computer device is programmed or configured to perform the steps of the multi-sensor-based online vehicle attitude warning method as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is programmed or configured to perform the multi-sensor-based online vehicle attitude warning method according to any one of claims 1 to 7.

Images