Motor vehicle steering wheel angle intelligent detection method and system based on machine learning

Abstract

The invention relates to the technical field of motor vehicle detection, in particular to a motor vehicle steering wheel angle intelligent detection method and system based on machine learning, and the method comprises the steps: constructing a motor vehicle steering wheel angle data analysis model through collected motor vehicle steering wheel angle sample data, and carrying out the preprocessing of the collected target motor vehicle steering wheel angle data; generating standard target motor vehicle steering wheel angle data, performing data analysis according to the motor vehicle steering wheel angle data analysis model and the standard target motor vehicle steering wheel angle data, generating target motor vehicle steering wheel angle data, and performing numerical comparison with a set angle threshold value; and generating target motor vehicle steering wheel angle calibration analysis result data according to the numerical value comparison result, if the target motor vehicle steering wheel angle calibration analysis result data is required to be calibrated, constructing target motor vehicle steering wheel angle real-time detection data, and executing target motor vehicle steering wheel calibration operation. And intelligent detection and accurate calibration of the steering wheel angle of the motor vehicle are realized.

Description

Technical Field

[0001] The present invention relates to the field of motor vehicle detection technology, and in particular to a motor vehicle steering wheel angle intelligent detection method and system based on machine learning. Background Art

[0002] With the development of the automobile industry and the improvement of people's living standards, more and more cars are appearing in the public eye. Among them, the stability of automobile operation is valued by people as an important indicator to measure automobile performance. The most intuitive test of automobile handling stability is the detection of automobile steering wheel rotation. However, in most cases nowadays, this process is still achieved manually, and the accuracy and reliability of the test results cannot be guaranteed.

[0003] The Chinese invention patent with publication number CN102878951B introduces an image-based vehicle steering wheel angle detection method and detection device. By marking on the steering wheel, obtaining the marked image, and identifying the marked image, the steering wheel rotation direction and steering wheel angle can be determined. Real-time detection of the steering wheel angle can be achieved, and the detection result is accurate. When the steering wheel rotation angle is greater than one circle, the steering wheel rotation angle can also be accurately obtained. However, it is difficult to ensure the accuracy of the steering wheel angle recognition result by only identifying the marked image, and it is also impossible to achieve accurate calibration of the steering wheel angle when the motor vehicle is unloaded. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] In order to solve the deficiencies in the background technology, the present invention provides a method and system for intelligent detection of the steering wheel angle of a motor vehicle based on machine learning, so as to realize intelligent detection and precise calibration of the steering wheel angle of the motor vehicle.

[0006] (II) Technical solution

[0007] A method for intelligently detecting a steering wheel angle of a motor vehicle based on machine learning comprises the following steps:

[0008] S1. Collecting sample data of steering wheel angle of motor vehicle;

[0009] S2. constructing a motor vehicle steering wheel angle data analysis model based on the motor vehicle steering wheel angle sample data;

[0010] S3, collecting steering wheel angle data of the target motor vehicle;

[0011] S4, performing data preprocessing on the target motor vehicle steering wheel angle data to generate standard target motor vehicle steering wheel angle data;

[0012] S5, performing target motor vehicle steering wheel angle data analysis and processing according to the motor vehicle steering wheel angle data analysis model and the standard target motor vehicle steering wheel angle data to generate target motor vehicle steering wheel angle data;

[0013] S6, setting an angle threshold, performing target motor vehicle steering wheel angle calibration analysis processing according to the angle threshold and the target motor vehicle steering wheel angle data, and generating target motor vehicle steering wheel angle calibration analysis result data according to the analysis result;

[0014] S7. If the target motor vehicle steering wheel angle calibration analysis result data indicates that calibration is required, target motor vehicle steering wheel angle real-time detection data is constructed according to the target motor vehicle steering wheel angle data and the target motor vehicle steering wheel angle calibration analysis result data, and a target motor vehicle steering wheel calibration operation is performed.

[0015] The present invention constructs a motor vehicle steering wheel angle data analysis model through collected motor vehicle steering wheel angle sample data, and performs data preprocessing on the collected target motor vehicle steering wheel angle data to generate standard target motor vehicle steering wheel angle data, performs data analysis processing based on the motor vehicle steering wheel angle data analysis model and the standard target motor vehicle steering wheel angle data to generate target motor vehicle steering wheel angle data, and performs target motor vehicle steering wheel angle calibration analysis processing with a set angle threshold, generates target motor vehicle steering wheel angle calibration analysis result data according to the analysis result, if the target motor vehicle steering wheel angle calibration analysis result data requires calibration, then constructs target motor vehicle steering wheel angle real-time detection data, and executes the target motor vehicle steering wheel calibration operation to realize intelligent detection and precise calibration of the motor vehicle steering wheel angle.

[0016] Preferably, the specific steps of collecting the sample data of the steering wheel angle of the motor vehicle are as follows:

[0017] S11. Collect steering wheel angle data of motor vehicles of different brands and models at different rotation angles online through the motor vehicle performance data management platform to generate a motor vehicle steering wheel angle sample data set A = {a1, a2, ..., a i ,…,a k}, where a i represents the i-th motor vehicle steering wheel angle sample data collected through the motor vehicle performance data management platform, k represents the total number of motor vehicle steering wheel angle sample data, and the angle data includes but is not limited to steering wheel angle data, steering wheel torque data, maximum free rotation inertia data and steering wheel rotation direction data.

[0018] A large amount of sample data is collected online through the motor vehicle performance data management platform to provide a data basis for the subsequent construction of a motor vehicle steering wheel angle data analysis model.

[0019] Preferably, the specific steps of constructing a motor vehicle steering wheel angle data analysis model based on the motor vehicle steering wheel angle sample data are as follows:

[0020] S21, setting a training sample ratio and a test sample ratio, dividing the motor vehicle steering wheel angle sample data set according to the training sample ratio and the test sample ratio, and obtaining a motor vehicle steering wheel angle training sample data set and a motor vehicle steering wheel angle test sample data set respectively;

[0021] S22, constructing an initial decision tree model, extracting n motor vehicle steering wheel angle training sample data from the motor vehicle steering wheel angle training sample data set by a method of extraction without replacement, forming a motor vehicle steering wheel angle training sample data subset with the n motor vehicle steering wheel angle training sample data, and training the initial decision tree model with the motor vehicle steering wheel angle training sample data subset to obtain a trained decision tree model;

[0022] The training of the initial decision tree model by using the motor vehicle steering wheel angle training sample data subset to obtain a trained decision tree model comprises the following steps:

[0023] S221, using the motor vehicle steering wheel angle training sample data in the motor vehicle steering wheel angle training sample data subset as the root node of the initial decision tree model, recorded as the first layer of the initial decision tree;

[0024] S222, using the CART algorithm to split the root node of the initial decision tree, generate child nodes, and then use the generated child nodes as root nodes to split again until no new child nodes are generated, and then end the splitting operation to obtain a trained decision tree model; the root node splitting process is as follows:

[0025] S2221. Setting a Gini coefficient threshold;

[0026] S2222, setting the motor vehicle steering wheel angle training sample data in the root node of the initial decision tree model to have N sample classification features, calculating the Gini coefficient of each sample classification feature, arranging the Gini coefficients of each sample classification feature from large to small, and selecting the maximum Gini coefficient to compare with the Gini coefficient threshold;

[0027] If the maximum value of the Gini coefficient is greater than the Gini coefficient threshold, the root node is further split according to the sample classification feature corresponding to the maximum value of the Gini coefficient;

[0028] If the Gini coefficient is less than or equal to the set Gini coefficient threshold, the root node splitting operation is terminated;

[0029] S23, repeat the operation in S22 until m trained decision tree models are obtained, and combine the m trained decision tree models to obtain a trained random forest model;

[0030] S24, evenly dividing the motor vehicle steering wheel angle test sample data set to generate a first motor vehicle steering wheel angle test sample data set and a second motor vehicle steering wheel angle test sample data set, setting a test accuracy threshold, inputting the first motor vehicle steering wheel angle test sample data set and the second motor vehicle steering wheel angle test sample data set into the trained random forest model for testing, respectively calculating the accuracy of the test results, and generating a first test accuracy and a second test accuracy;

[0031] S25. Use the mean formula to perform average numerical measurement processing on the first test accuracy and the second test accuracy to obtain a final test accuracy. When the final test accuracy is greater than or equal to the test accuracy threshold, use the trained random forest model as the motor vehicle steering wheel angle data analysis model; otherwise, use the Empire Competition Optimization Algorithm to tune the parameters of the trained random forest model, return to S22, and retrain until the final test accuracy is greater than or equal to the test accuracy threshold.

[0032] The parameter tuning of the trained random forest model by the imperial competition optimization algorithm includes the following steps:

[0033] S251, construct national population, set national population size to N, assimilation coefficient to χ, current iteration number to t, maximum iteration number to t max And the dimension of the random forest model parameter search space is P;

[0034] Randomly generate N groups of random forest model parameters in the random forest model parameter search space, each group of random forest model parameters corresponds to a country individual in the country population, and initialize the initial position set of the country population to X = {X1, X2, ..., X i ,…,X N}, where X i represents the initial position of the individual of the ith country in the country population;

[0035] S252. Calculate the power value of each country in the country population, arrange the country individuals in the country population from large to small according to the power value, select the M country individuals with the highest power value as imperialist country individuals, and the remaining country individuals in the country population as colonial country individuals, and randomly distribute the colonial country individuals to the imperialist country individuals according to the size ratio of the imperialist country individual power value, forming M empires; the power value calculation formula is as follows:

[0036]

[0037] Among them, Q i represents the power value of the individual in the i-th country in the national population, x i It represents the accuracy of the test results of the random forest model obtained by tuning the parameters of the trained random forest model by the random forest model parameters corresponding to the i-th country individual in the country population, and φ represents the correction value;

[0038] S253. The imperialist country individuals in each empire will assimilate the colonial country individuals in the empire in the parameter search space of the random forest model, causing the colonial country individuals to move to the imperialist country individuals in the empire for position update; the position update formula is as follows:

[0039]

[0040] in, and Denote the position of the individual of the i-th colony country in the country population after and before the position update, respectively. i V is the distance between the individual of the ith colonial country in the country population and the individual of the imperialist country in its empire. i represents the direction vector of the i-th colonial country individual in the country population pointing to the imperialist country individual in its empire, and U(0,χ·D) means that it obeys the uniform distribution between [0,χ·D];

[0041] S254. After assimilation, the individuals of the colonized countries in each empire will perform revolution operations in the random forest model parameter search space according to the set revolution rate, and update their positions by random movement;

[0042] S255. Calculate the power value of each colony country after the position update. If the power value of a colony country is greater than the power value of an imperialist country in the empire where it is located, the colony country individual becomes a new imperialist country individual, and the original imperialist country individual becomes a colony country individual.

[0043] S256. Calculate the total power of each empire. The empire with the lowest total power will give up the colony country with the lowest power in the empire. Other empires will occupy the colony country according to probability. If there is no colony country in the empire, the empire will be abolished, and the imperialist country of the empire will be allocated to the empire with the highest total power as a colony country. The probability calculation formula is as follows:

[0044]

[0045] Among them, P i represents the probability that the i-th empire occupies the colonial country with the lowest power value in the empire with the lowest total power value, represents the total power value of the i-th empire;

[0046] S257: Determine whether the current number of iterations t is less than the maximum number of iterations t max , if the current number of iterations t is less than the maximum number of iterations t max , then the current iteration number t is increased by 1, and the process returns to S253; otherwise, the country individual with the highest power value in the country population is taken as the global optimal solution, and the parameters of the trained random forest model are adjusted to the random forest model parameters corresponding to the global optimal solution.

[0047] A motor vehicle steering wheel angle data analysis model is constructed based on the motor vehicle steering wheel angle sample data, which provides a good tool for realizing accurate analysis of motor vehicle steering wheel angle data, ensures the reliability of the motor vehicle steering wheel angle data analysis results, and efficiently, accurately and quickly analyzes the angle data of the target motor vehicle steering wheel; the parameters of the trained random forest model are tuned through the empire competition optimization algorithm, which improves the prediction accuracy of the acquired motor vehicle steering wheel angle data analysis model, improves the speed and efficiency of the parameter optimization process, and can effectively avoid the parameter optimization results from falling into the local optimal solution.

[0048] Preferably, the specific steps of collecting the steering wheel angle data of the target motor vehicle are as follows:

[0049] S31. Collect steering wheel angle data of the target motor vehicle in an unloaded state in real time through different sensors to generate steering wheel angle data B of the target motor vehicle. The different sensors include but are not limited to angle sensors, torque sensors, and gyroscopes.

[0050] Preferably, the specific steps of preprocessing the target motor vehicle steering wheel angle data to generate standard target motor vehicle steering wheel angle data are as follows:

[0051] S41, performing target motor vehicle steering wheel angle data filtering and noise reduction processing on the target motor vehicle steering wheel angle data B by using an exponential sliding average filtering method to generate standard target motor vehicle steering wheel angle data C.

[0052] The exponential sliding average filtering method is used to filter and reduce the noise of the target motor vehicle steering wheel angle data, which effectively avoids the interference caused by noise in the acquired data, improves the quality and accuracy of the acquired data, and ensures that the acquired data can accurately and truly reflect the real-time situation of the target motor vehicle steering wheel angle.

[0053] Preferably, the specific steps of performing target motor vehicle steering wheel angle data analysis and processing based on the motor vehicle steering wheel angle data analysis model and the standard target motor vehicle steering wheel angle data to generate target motor vehicle steering wheel angle data are as follows:

[0054] S51, inputting the standard target motor vehicle steering wheel angle data C into the motor vehicle steering wheel angle data analysis model to perform target motor vehicle steering wheel angle data analysis and processing, and generating target motor vehicle steering wheel angle data D jiaodu .

[0055] Preferably, an angle threshold is set, and target motor vehicle steering wheel angle calibration analysis processing is performed according to the angle threshold and the target motor vehicle steering wheel angle data, and target motor vehicle steering wheel angle calibration analysis result data is generated according to the analysis result. The specific steps are as follows:

[0056] S61, setting the angle threshold to θ;

[0057] S62, the angle threshold θ and the target vehicle steering wheel angle data D jiaodu Perform numerical comparison and construct the target motor vehicle steering wheel angle calibration analysis result data E according to the numerical comparison results fenxi ;

[0058] When the angle threshold θ is greater than or equal to the target vehicle steering wheel angle data D jiaodu When , it indicates that the target motor vehicle steering wheel angle is in a normal state, and the target motor vehicle steering wheel angle calibration analysis result data E is output. fenxi No calibration is required;

[0059] When the angle threshold θ is less than the target vehicle steering wheel angle data D jiaodu When , it indicates that the steering wheel angle of the target motor vehicle is in an abnormal state, and the target motor vehicle steering wheel angle calibration analysis result data E is output. fenxi Calibration is required.

[0060] By setting an angle threshold and comparing the value with the steering wheel angle data of the target motor vehicle, it is accurately analyzed whether the steering wheel angle of the target motor vehicle needs to be calibrated.

[0061] Preferably, if the target motor vehicle steering wheel angle calibration analysis result data is in need of calibration, then the target motor vehicle steering wheel angle real-time detection data is constructed according to the target motor vehicle steering wheel angle data and the target motor vehicle steering wheel angle calibration analysis result data, and the specific steps of performing the target motor vehicle steering wheel calibration operation are as follows:

[0062] S71: If the target motor vehicle steering wheel angle calibration analysis result data indicates that calibration is required, the target motor vehicle steering wheel angle data D jiaodu and the target motor vehicle steering wheel angle calibration analysis result data E fenxi Combine the data to construct the real-time detection data of the steering wheel angle of the target motor vehicle F = (D jiaodu ,E fenxi );

[0063] S72, transmitting the real-time detection data F of the steering wheel angle of the target motor vehicle to the target motor vehicle through the wireless communication network. jiaodu ,E fenxi ) is pushed to the motor vehicle steering wheel angle real-time detection platform, and the target motor vehicle steering wheel calibration operation is performed.

[0064] The present invention also discloses a motor vehicle steering wheel angle intelligent detection system based on machine learning, comprising a motor vehicle steering wheel angle sample data acquisition module, a motor vehicle steering wheel angle data analysis model construction module, a target motor vehicle steering wheel angle data acquisition module, a target motor vehicle steering wheel angle data preprocessing module, a target motor vehicle steering wheel angle data analysis module, a target motor vehicle steering wheel angle calibration analysis module and a target motor vehicle steering wheel angle real-time detection data construction module;

[0065] The motor vehicle steering wheel angle sample data collection module collects the steering wheel angle data of motor vehicles of different brands and models at different rotation angles online through the motor vehicle performance data management platform to generate motor vehicle steering wheel angle sample data;

[0066] The motor vehicle steering wheel angle data analysis model construction module constructs a motor vehicle steering wheel angle data analysis model based on the motor vehicle steering wheel angle sample data;

[0067] The target motor vehicle steering wheel angle data acquisition module collects the steering wheel angle data of the target motor vehicle in an unloaded state in real time through different sensors to generate the target motor vehicle steering wheel angle data;

[0068] The target motor vehicle steering wheel angle data preprocessing module uses an exponential sliding average filtering method to perform target motor vehicle steering wheel angle data filtering and noise reduction processing on the target motor vehicle steering wheel angle data to generate standard target motor vehicle steering wheel angle data;

[0069] The target motor vehicle steering wheel angle data analysis module inputs the standard target motor vehicle steering wheel angle data into the motor vehicle steering wheel angle data analysis model to perform target motor vehicle steering wheel angle data analysis processing to generate target motor vehicle steering wheel angle data;

[0070] The target motor vehicle steering wheel angle calibration analysis module sets an angle threshold, compares the angle threshold with the target motor vehicle steering wheel angle data, and constructs target motor vehicle steering wheel angle calibration analysis result data according to the numerical comparison result;

[0071] When the target motor vehicle steering wheel angle calibration analysis result data requires calibration, the target motor vehicle steering wheel angle real-time detection data construction module combines the target motor vehicle steering wheel angle data and the target motor vehicle steering wheel angle calibration analysis result data to construct the target motor vehicle steering wheel angle real-time detection data, which is then pushed to the motor vehicle steering wheel angle real-time detection platform via the wireless communication network, and the target motor vehicle steering wheel calibration operation is performed.

[0072] (III) Beneficial effects

[0073] 1. The present invention constructs a motor vehicle steering wheel angle data analysis model through collected motor vehicle steering wheel angle sample data, performs data preprocessing on the collected target motor vehicle steering wheel angle data, generates standard target motor vehicle steering wheel angle data, performs data analysis processing according to the motor vehicle steering wheel angle data analysis model and the standard target motor vehicle steering wheel angle data, generates target motor vehicle steering wheel angle data, performs target motor vehicle steering wheel angle calibration analysis processing with a set angle threshold, generates target motor vehicle steering wheel angle calibration analysis result data according to the analysis result, if the target motor vehicle steering wheel angle calibration analysis result data is that calibration is required, constructs target motor vehicle steering wheel angle real-time detection data, and performs target motor vehicle steering wheel calibration operation, thereby realizing intelligent detection and precise calibration of the motor vehicle steering wheel angle;

[0074] 2. A motor vehicle steering wheel angle data analysis model is constructed based on the motor vehicle steering wheel angle sample data, which provides a good tool for accurately analyzing the motor vehicle steering wheel angle data, ensures the reliability of the motor vehicle steering wheel angle data analysis results, and efficiently, accurately and quickly analyzes the angle data of the target motor vehicle steering wheel;

[0075] 3. The parameters of the trained random forest model are optimized through the imperial competition optimization algorithm, which improves the prediction accuracy of the acquired motor vehicle steering wheel angle data analysis model, improves the speed and efficiency of the parameter optimization process, and effectively avoids the parameter optimization results from falling into the local optimal solution;

[0076] 4. The exponential sliding average filtering method is used to filter and reduce the noise of the steering wheel angle data of the target motor vehicle, which effectively avoids the interference caused by noise in the acquired data, improves the quality and accuracy of the acquired data, and ensures that the acquired data can accurately and truly reflect the real-time situation of the steering wheel angle of the target motor vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0078] Figure 1 A flow chart of a method for intelligently detecting a steering wheel angle of a motor vehicle based on machine learning provided by the present invention;

[0079] Figure 2 A module schematic diagram of a motor vehicle steering wheel angle intelligent detection system based on machine learning provided by the present invention. DETAILED DESCRIPTION

[0080] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0082] Embodiment 1 is as follows:

[0083] See also Figure 1 , a method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning, comprising the following steps:

[0084] S1. Collecting sample data of steering wheel angle of motor vehicle;

[0085] S11. Collect steering wheel angle data of motor vehicles of different brands and models at different rotation angles online through the motor vehicle performance data management platform to generate a motor vehicle steering wheel angle sample data set A = {a1, a2, ..., a i ,…,a k}, where a irepresents the i-th motor vehicle steering wheel angle sample data collected through the motor vehicle performance data management platform, k represents the total number of motor vehicle steering wheel angle sample data, and the angle data includes but is not limited to steering wheel angle data, steering wheel torque data, maximum free rotation inertia data and steering wheel rotation direction data.

[0086] S2. constructing a motor vehicle steering wheel angle data analysis model based on the motor vehicle steering wheel angle sample data;

[0087] S21, setting a training sample ratio and a test sample ratio, dividing the motor vehicle steering wheel angle sample data set according to the training sample ratio and the test sample ratio, and obtaining a motor vehicle steering wheel angle training sample data set and a motor vehicle steering wheel angle test sample data set respectively;

[0088] S22, constructing an initial decision tree model, extracting n motor vehicle steering wheel angle training sample data from the motor vehicle steering wheel angle training sample data set by a method of extraction without replacement, forming a motor vehicle steering wheel angle training sample data subset with the n motor vehicle steering wheel angle training sample data, and training the initial decision tree model with the motor vehicle steering wheel angle training sample data subset to obtain a trained decision tree model;

[0089] The training of the initial decision tree model by using the motor vehicle steering wheel angle training sample data subset to obtain a trained decision tree model comprises the following steps:

[0090] S221, using the motor vehicle steering wheel angle training sample data in the motor vehicle steering wheel angle training sample data subset as the root node of the initial decision tree model, recorded as the first layer of the initial decision tree;

[0091] S222, using the CART algorithm to split the root node of the initial decision tree, generate child nodes, and then use the generated child nodes as root nodes to split again until no new child nodes are generated, and then end the splitting operation to obtain a trained decision tree model; the root node splitting process is as follows:

[0092] S2221. Setting a Gini coefficient threshold;

[0093] S2222, setting the motor vehicle steering wheel angle training sample data in the root node of the initial decision tree model to have N sample classification features, calculating the Gini coefficient of each sample classification feature, arranging the Gini coefficients of each sample classification feature from large to small, and selecting the maximum Gini coefficient to compare with the Gini coefficient threshold;

[0094] If the maximum value of the Gini coefficient is greater than the Gini coefficient threshold, the root node is further split according to the sample classification feature corresponding to the maximum value of the Gini coefficient;

[0095] If the Gini coefficient is less than or equal to the set Gini coefficient threshold, the root node splitting operation is terminated;

[0096] S23, repeat the operation in S22 until m trained decision tree models are obtained, and combine the m trained decision tree models to obtain a trained random forest model;

[0097] S24, evenly dividing the motor vehicle steering wheel angle test sample data set to generate a first motor vehicle steering wheel angle test sample data set and a second motor vehicle steering wheel angle test sample data set, setting a test accuracy threshold, inputting the first motor vehicle steering wheel angle test sample data set and the second motor vehicle steering wheel angle test sample data set into the trained random forest model for testing, respectively calculating the accuracy of the test results, and generating a first test accuracy and a second test accuracy;

[0098] S25. Use the mean formula to perform average numerical measurement processing on the first test accuracy and the second test accuracy to obtain a final test accuracy. When the final test accuracy is greater than or equal to the test accuracy threshold, use the trained random forest model as the motor vehicle steering wheel angle data analysis model; otherwise, use the Empire Competition Optimization Algorithm to tune the parameters of the trained random forest model, return to S22, and retrain until the final test accuracy is greater than or equal to the test accuracy threshold.

[0099] The parameter tuning of the trained random forest model by the imperial competition optimization algorithm includes the following steps:

[0100] S251, construct national population, set national population size to N, assimilation coefficient to χ, current iteration number to t, maximum iteration number to t max And the dimension of the random forest model parameter search space is P;

[0101] Randomly generate N groups of random forest model parameters in the random forest model parameter search space, each group of random forest model parameters corresponds to a country individual in the country population, and initialize the initial position set of the country population to X = {X1, X2, ..., X i ,…,X N}, where X i represents the initial position of the individual of the ith country in the country population;

[0102] S252. Calculate the power value of each country in the country population, arrange the country individuals in the country population from large to small according to the power value, select the M country individuals with the highest power value as imperialist country individuals, and the remaining country individuals in the country population as colonial country individuals, and randomly distribute the colonial country individuals to the imperialist country individuals according to the size ratio of the imperialist country individual power value, forming M empires; the power value calculation formula is as follows:

[0103]

[0104] Among them, Q i represents the power value of the individual in the i-th country in the national population, x i It represents the accuracy of the test results of the random forest model obtained by tuning the parameters of the trained random forest model by the random forest model parameters corresponding to the i-th country individual in the country population, and φ represents the correction value;

[0105] S253. The imperialist country individuals in each empire will assimilate the colonial country individuals in the empire in the parameter search space of the random forest model, causing the colonial country individuals to move to the imperialist country individuals in the empire for position update; the position update formula is as follows:

[0106]

[0107] in, and Denote the position of the individual of the i-th colony country in the country population after and before the position update, respectively. i V is the distance between the individual of the ith colonial country in the country population and the individual of the imperialist country in its empire. i represents the direction vector of the i-th colonial country individual in the country population pointing to the imperialist country individual in its empire, and U(0,χ·D) means that it obeys the uniform distribution between [0,χ·D];

[0108] S254. After assimilation, the individuals of the colonized countries in each empire will perform revolution operations in the random forest model parameter search space according to the set revolution rate, and update their positions by random movement;

[0109] S255. Calculate the power value of each colony country after the position update. If the power value of a colony country is greater than the power value of an imperialist country in the empire where it is located, the colony country individual becomes a new imperialist country individual, and the original imperialist country individual becomes a colony country individual.

[0110] S256. Calculate the total power of each empire. The empire with the lowest total power will give up the colony country with the lowest power in the empire. Other empires will occupy the colony country according to probability. If there is no colony country in the empire, the empire will be abolished, and the imperialist country of the empire will be allocated to the empire with the highest total power as a colony country. The probability calculation formula is as follows:

[0111]

[0112] Among them, P i represents the probability that the i-th empire occupies the colony with the lowest power value in the empire with the lowest total power value, Q i sum represents the total power value of the i-th empire;

[0113] S257: Determine whether the current number of iterations t is less than the maximum number of iterations t max , if the current number of iterations t is less than the maximum number of iterations t max , then the current iteration number t is increased by 1, and the process returns to S253; otherwise, the country individual with the highest power value in the country population is taken as the global optimal solution, and the parameters of the trained random forest model are adjusted to the random forest model parameters corresponding to the global optimal solution.

[0114] S3, collecting steering wheel angle data of the target motor vehicle;

[0115] S31. Collect steering wheel angle data of the target motor vehicle in an unloaded state in real time through different sensors to generate steering wheel angle data B of the target motor vehicle. The different sensors include but are not limited to angle sensors, torque sensors, and gyroscopes.

[0116] S4, performing data preprocessing on the target motor vehicle steering wheel angle data to generate standard target motor vehicle steering wheel angle data;

[0117] S41, performing target motor vehicle steering wheel angle data filtering and noise reduction processing on the target motor vehicle steering wheel angle data B by using an exponential sliding average filtering method to generate standard target motor vehicle steering wheel angle data C.

[0118] S5, performing target motor vehicle steering wheel angle data analysis and processing according to the motor vehicle steering wheel angle data analysis model and the standard target motor vehicle steering wheel angle data to generate target motor vehicle steering wheel angle data;

[0119] S51, inputting the standard target motor vehicle steering wheel angle data C into the motor vehicle steering wheel angle data analysis model to perform target motor vehicle steering wheel angle data analysis and processing, and generating target motor vehicle steering wheel angle data D jiaodu .

[0120] S6, setting an angle threshold, performing target motor vehicle steering wheel angle calibration analysis processing according to the angle threshold and the target motor vehicle steering wheel angle data, and generating target motor vehicle steering wheel angle calibration analysis result data according to the analysis result;

[0121] S61, setting the angle threshold to θ;

[0122] S62, the angle threshold θ and the target vehicle steering wheel angle data D jiaodu Perform numerical comparison and construct the target motor vehicle steering wheel angle calibration analysis result data E according to the numerical comparison results fenxi ;

[0123] When the angle threshold θ is greater than or equal to the target vehicle steering wheel angle data D jiaodu When , it indicates that the target motor vehicle steering wheel angle is in a normal state, and the target motor vehicle steering wheel angle calibration analysis result data E is output. fenxi No calibration is required;

[0124] When the angle threshold θ is less than the target vehicle steering wheel angle data D jiaodu When , it indicates that the steering wheel angle of the target motor vehicle is in an abnormal state, and the target motor vehicle steering wheel angle calibration analysis result data E is output. fenxi Calibration is required.

[0125] S7, if the target motor vehicle steering wheel angle calibration analysis result data indicates that calibration is required, constructing target motor vehicle steering wheel angle real-time detection data according to the target motor vehicle steering wheel angle data and the target motor vehicle steering wheel angle calibration analysis result data, and performing a target motor vehicle steering wheel calibration operation;

[0126] S71: If the target motor vehicle steering wheel angle calibration analysis result data indicates that calibration is required, the target motor vehicle steering wheel angle data D jiaodu and the target motor vehicle steering wheel angle calibration analysis result data E fenxi Combine the data to construct the real-time detection data of the steering wheel angle of the target motor vehicle F = (D jiaodu ,E fenxi );

[0127] S72, transmitting the real-time detection data F of the steering wheel angle of the target motor vehicle to the target motor vehicle through the wireless communication network. jiaodu ,E fenxi ) is pushed to the motor vehicle steering wheel angle real-time detection platform, and the target motor vehicle steering wheel calibration operation is performed.

[0128] Embodiment 2 is as follows:

[0129] See also Figure 2 , a motor vehicle steering wheel angle intelligent detection system based on machine learning, comprising a motor vehicle steering wheel angle sample data acquisition module, a motor vehicle steering wheel angle data analysis model construction module, a target motor vehicle steering wheel angle data acquisition module, a target motor vehicle steering wheel angle data preprocessing module, a target motor vehicle steering wheel angle data analysis module, a target motor vehicle steering wheel angle calibration analysis module and a target motor vehicle steering wheel angle real-time detection data construction module;

[0130] The motor vehicle steering wheel angle sample data collection module collects the steering wheel angle data of motor vehicles of different brands and models at different rotation angles online through the motor vehicle performance data management platform to generate motor vehicle steering wheel angle sample data;

[0131] The motor vehicle steering wheel angle data analysis model construction module constructs a motor vehicle steering wheel angle data analysis model based on the motor vehicle steering wheel angle sample data;

[0132] The target motor vehicle steering wheel angle data acquisition module collects the steering wheel angle data of the target motor vehicle in an unloaded state in real time through different sensors to generate the target motor vehicle steering wheel angle data;

[0133] The target motor vehicle steering wheel angle data preprocessing module uses an exponential sliding average filtering method to perform target motor vehicle steering wheel angle data filtering and noise reduction processing on the target motor vehicle steering wheel angle data to generate standard target motor vehicle steering wheel angle data;

[0134] The target motor vehicle steering wheel angle data analysis module inputs the standard target motor vehicle steering wheel angle data into the motor vehicle steering wheel angle data analysis model to perform target motor vehicle steering wheel angle data analysis processing to generate target motor vehicle steering wheel angle data;

[0135] The target motor vehicle steering wheel angle calibration analysis module sets an angle threshold, compares the angle threshold with the target motor vehicle steering wheel angle data, and constructs target motor vehicle steering wheel angle calibration analysis result data according to the numerical comparison result;

[0136] When the target motor vehicle steering wheel angle calibration analysis result data requires calibration, the target motor vehicle steering wheel angle real-time detection data construction module combines the target motor vehicle steering wheel angle data and the target motor vehicle steering wheel angle calibration analysis result data to construct the target motor vehicle steering wheel angle real-time detection data, which is then pushed to the motor vehicle steering wheel angle real-time detection platform via the wireless communication network, and the target motor vehicle steering wheel calibration operation is performed.

[0137] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0138] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning, characterized in that: The steps include: S1. Collecting sample data of steering wheel angle of motor vehicle; S2. constructing a motor vehicle steering wheel angle data analysis model based on the motor vehicle steering wheel angle sample data; S3, collecting steering wheel angle data of the target motor vehicle; S4, performing data preprocessing on the target motor vehicle steering wheel angle data to generate standard target motor vehicle steering wheel angle data; S5, performing target motor vehicle steering wheel angle data analysis and processing according to the motor vehicle steering wheel angle data analysis model and the standard target motor vehicle steering wheel angle data to generate target motor vehicle steering wheel angle data; S6, setting an angle threshold, performing target motor vehicle steering wheel angle calibration analysis processing according to the angle threshold and the target motor vehicle steering wheel angle data, and generating target motor vehicle steering wheel angle calibration analysis result data according to the analysis result; S7. If the target motor vehicle steering wheel angle calibration analysis result data indicates that calibration is required, target motor vehicle steering wheel angle real-time detection data is constructed according to the target motor vehicle steering wheel angle data and the target motor vehicle steering wheel angle calibration analysis result data, and a target motor vehicle steering wheel calibration operation is performed.

2. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 1, characterized in that: The S1 comprises the following steps: S11. Collect steering wheel angle data of motor vehicles of different brands and models at different rotation angles online through a motor vehicle performance data management platform to generate a motor vehicle steering wheel angle sample data set.

3. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 2, characterized in that: The S2 comprises the following steps: S21, setting a training sample ratio and a test sample ratio, dividing the motor vehicle steering wheel angle sample data set according to the training sample ratio and the test sample ratio, and obtaining a motor vehicle steering wheel angle training sample data set and a motor vehicle steering wheel angle test sample data set respectively; S22, constructing an initial decision tree model, extracting n motor vehicle steering wheel angle training sample data from the motor vehicle steering wheel angle training sample data set by a method of extraction without replacement, forming a motor vehicle steering wheel angle training sample data subset with the n motor vehicle steering wheel angle training sample data, and training the initial decision tree model with the motor vehicle steering wheel angle training sample data subset to obtain a trained decision tree model; The training of the initial decision tree model by using the motor vehicle steering wheel angle training sample data subset to obtain a trained decision tree model comprises the following steps: S221, using the motor vehicle steering wheel angle training sample data in the motor vehicle steering wheel angle training sample data subset as the root node of the initial decision tree model, recorded as the first layer of the initial decision tree; S222, using the CART algorithm to split the root node of the initial decision tree to generate child nodes, and then using the generated child nodes as root nodes to split again until no new child nodes are generated, and then ending the splitting operation to obtain a trained decision tree model; S23, repeat the operation in S22 until m trained decision tree models are obtained, and combine the m trained decision tree models to obtain a trained random forest model; S24, evenly dividing the motor vehicle steering wheel angle test sample data set to generate a first motor vehicle steering wheel angle test sample data set and a second motor vehicle steering wheel angle test sample data set, setting a test accuracy threshold, inputting the first motor vehicle steering wheel angle test sample data set and the second motor vehicle steering wheel angle test sample data set into the trained random forest model for testing, respectively calculating the accuracy of the test results, and generating a first test accuracy and a second test accuracy; S25. Use the mean formula to perform average numerical measurement processing on the first test accuracy and the second test accuracy to obtain a final test accuracy. When the final test accuracy is greater than or equal to the test accuracy threshold, the trained random forest model is used as a motor vehicle steering wheel angle data analysis model; otherwise, the trained random forest model is parameter tuned through the imperial competition optimization algorithm, and returns to S22 for retraining until the final test accuracy is greater than or equal to the test accuracy threshold.

4. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 3, characterized in that: In S25, the parameter tuning of the trained random forest model by the imperial competition optimization algorithm includes the following steps: S251, construct a national population, set the national population size to N, the current number of iterations to t, and the maximum number of iterations to t max And the dimension of the random forest model parameter search space is P; Randomly generating N groups of random forest model parameters in the random forest model parameter search space, each group of random forest model parameters corresponding to a national individual in the national population; S252. Calculate the power value of each country in the country population, arrange the country individuals in the country population from large to small according to the power value, select the M country individuals with the highest power value as imperialist country individuals, and the remaining country individuals in the country population as colonial country individuals, and randomly distribute the colonial country individuals to the imperialist country individuals according to the ratio of the power value of the imperialist country individuals, so as to form M empires; S253, the imperialist country individuals in each empire will assimilate the colonial country individuals in the empire in which they are located in the parameter search space of the random forest model, causing the colonial country individuals to move toward the imperialist country individuals in the empire in which they are located to update their positions; S254. After assimilation, the individuals of the colonized countries in each empire will perform revolution operations in the random forest model parameter search space according to the set revolution rate, and update their positions by random movement; S255. Calculate the power value of each colony country after the position update. If the power value of a colony country is greater than the power value of an imperialist country in the empire where it is located, the colony country individual becomes a new imperialist country individual, and the original imperialist country individual becomes a colony country individual. S256. Calculate the total power of each empire. The empire with the lowest total power will give up the colony country with the lowest power in the empire. Other empires will occupy the colony country according to probability. If there is no colony country in the empire, the empire will be abolished, and the imperialist country of the empire will be allocated to the empire with the highest total power as a colony country. S257, determine whether t is less than t max , if t is less than t max , then t is increased by 1, and the process returns to S253; otherwise, the country individual with the highest power value in the country population is taken as the global optimal solution, and the parameters of the trained random forest model are adjusted to the random forest model parameters corresponding to the global optimal solution.

5. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 4, characterized in that: The S3 comprises the following steps: S31. Using different sensors to collect steering wheel angle data of a target motor vehicle in an unloaded state in real time, and generate steering wheel angle data B of the target motor vehicle.

6. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 5, characterized in that: The S4 comprises the following steps: S41, using an exponential sliding average filtering method to perform filtering and noise reduction processing on the target motor vehicle steering wheel angle data B to generate standard target motor vehicle steering wheel angle data C.

7. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 6, characterized in that: The S5 comprises the following steps: S51, inputting the C into the motor vehicle steering wheel angle data analysis model to perform target motor vehicle steering wheel angle data analysis and processing, and generating target motor vehicle steering wheel angle data D jiaodu .

8. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 7, characterized in that: The S6 comprises the following steps: S61, setting the angle threshold to θ; S62, the θ and the D jiaodu Perform numerical comparison and construct the target motor vehicle steering wheel angle calibration analysis result data E according to the numerical comparison results fenxi ; When the θ is greater than or equal to the D jiaodu When , it indicates that the target motor vehicle steering wheel angle is in a normal state, and the target motor vehicle steering wheel angle calibration analysis result data E is output. fenxi No calibration is required; When the θ is less than the D jiaodu When , it indicates that the steering wheel angle of the target motor vehicle is in an abnormal state, and the target motor vehicle steering wheel angle calibration analysis result data E is output. fenxi Calibration is required.

9. The method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning according to claim 8, characterized in that: The S7 comprises the following steps: S71, if the target motor vehicle steering wheel angle calibration analysis result data is required to calibrate, then the D jiaodu and the E fenxi Combine the data to construct the real-time detection data of the steering wheel angle of the target motor vehicle F = (D jiaodu ,E fenxi ); S72, through the wireless communication network, the F=(D jiaodu ,E fenxi ) is pushed to the motor vehicle steering wheel angle real-time detection platform, and the target motor vehicle steering wheel calibration operation is performed.

10. A system for implementing the method for intelligent detection of steering wheel angle of a motor vehicle based on machine learning as described in any one of claims 1 to 9.

Images