An intelligent control system of motor control machine based on vehicle behavior data

Abstract

The application discloses a motor control machine intelligent control system based on vehicle behavior data, and relates to the technical field of motor automation control, and comprises a data acquisition module, which is used for acquiring environmental data, vehicle motion data and driver state data in the vehicle driving process; a weather adaptation model construction module, which is used for evaluating vehicle performance change scores of drivers under different weather conditions, establishing a driver performance evaluation strategy, evaluating driver proficiency under different weather conditions according to the vehicle performance, and outputting driver trustworthiness; and a self-adaptive parameter adjustment module, which is used for constructing a self-adaptive motor control model, performing weight distribution on control input according to the driver trustworthiness, and outputting control parameters. The application solves the motor control error caused by the personal habits of drivers under different weather conditions.

Description

Technical Field

[0001] This invention relates to the field of motor automation control technology, specifically to an intelligent control system for motor control based on vehicle behavior data. Background Technology

[0002] Traditional motor control methods typically pre-design a globally consistent set of weights or gains, which cannot dynamically adjust to different drivers' driving styles, experience levels, and physiological and psychological states such as fatigue and attention fluctuations during driving. This "one-size-fits-all" control strategy not only fails to meet the precise control needs of skilled drivers in dynamic scenarios, but also cannot intervene in a timely manner when the driver is not in good condition, leading to safety hazards.

[0003] While complex control algorithms perform well in simulations and small-scale tests, their high computational cost and dependence on model accuracy make them difficult to meet the millisecond-level real-time control requirements of resource-constrained automotive ECUs / MCUs. To reduce computational overhead, many solutions simplify the algorithm or reduce the prediction time domain, thereby sacrificing control accuracy and robustness, making them difficult to promote and apply on actual roads.

[0004] With the rise of the "human-machine shared control" concept, how to dynamically allocate control between the driver and the system to ensure safety without depriving the driver of initiative has become an urgent problem to be solved. Existing research mostly focuses on theoretical models based on confidence levels or simple weighting, lacking a comprehensive consideration of the driver's real-time state, preferences, and historical performance, and also lacking a unified interface and process to tightly integrate trust levels, weather perception, and control strategies.

[0005] Different drivers have varying levels of proficiency and familiarity with different weather conditions. A driver may be more familiar with snowy roads but less so with rainy roads. Therefore, different drivers need to respond differently to different weather types and conditions. In particular, how to quickly and accurately switch control in emergency situations to ensure driving safety still requires further exploration.

[0006] Therefore, the present invention provides an intelligent control system for motor control based on vehicle behavior data. Summary of the Invention

[0007] The purpose of this invention is to provide an intelligent control system for motor control based on vehicle behavior data, so as to solve the existing problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an intelligent control system for motor control based on vehicle behavior data, comprising:

[0009] The data acquisition module is used to collect environmental data, vehicle motion data, and driver status data during vehicle operation.

[0010] The weather adaptation model building module is used to evaluate the vehicle performance changes of drivers and vehicles under different weather conditions, establish a driver performance evaluation strategy, evaluate driver proficiency under different weather conditions based on vehicle performance changes, and output driver trust level.

[0011] The adaptive parameter adjustment module is used to construct an adaptive motor control model, which distributes the control inputs with weights according to the driver's trust level and outputs control parameters.

[0012] A further improvement of this invention is that the data acquisition module includes environmental data, vehicle motion data, and driver status data; the environmental data includes camera images, millimeter-wave radar point clouds, and lidar point clouds; the vehicle motion data includes vehicle speed, acceleration, lateral displacement, wheel speed, and steering angle; the driver status data includes eye tracking, facial expressions, and heart rate monitoring; all data are aligned using timestamps, and noise is eliminated using Kalman filtering.

[0013] A further improvement of this invention is that the weather adaptation model construction module includes a performance index evaluation unit, a proficiency assessment unit, and a driver trust unit. The performance index evaluation unit is used to obtain a vehicle performance change score based on the absolute value of the difference between the weighted score of the vehicle trajectory error, speed deviation, slip rate, road friction, and smoothness after normalization and 1. The performance index weights are updated based on a weather-weight mapping model, specifically: a weather perception feature vector E is constructed, and based on a small fully connected neural network, the combination of performance index weights that maximizes the vehicle performance change score is searched for as a label from historical operating data. With the goal of minimizing the MSE between the regression output and the label, an unnormalized performance index weight sequence vector is output, and the final performance index weights are obtained through Softmax.

[0014] A further improvement of this invention is that the performance index evaluation unit further includes, after each evaluation window ends, calculating the vehicle performance change score Cvp for the current window, setting the target performance score as Pcvp, and deriving the performance index error. ,when When the error exceeds the preset performance index threshold, perform gradient updates on the weather-weighted mapping model. ,in Indicates the learning rate. This represents the sequence of performance metric weights for the k-th iteration, with each performance metric weight set as follows: ,when The iteration stops when the change value is less than the set threshold.

[0015] A further improvement of this invention is that the proficiency assessment unit is equipped with a driver performance evaluation strategy, specifically including: creating a weather type vector W, scoring based on vehicle performance changes, and constructing a long-term proficiency model of the driver under various weather conditions, which is expressed as follows: for the k-th iteration... , This represents the driver i's proficiency under weather type w, and the initial proficiency of the driver is preset for each weather type. , This indicates the score for the changes in vehicle performance.

[0016] A further improvement of this invention is that the driver trust unit is used to determine driver trust based on the driver's current state and historical state; specifically, it includes: obtaining the average historical driver state score based on historical driver state data for the current weather type, using it as the driver's historical state baseline dsh; and calculating the driver's current state score dsc for time intervals t within a past time window T in the same way; calculating the difference between the driver's current state score and the historical state baseline as the state baseline correction factor Δds, thus obtaining the driver trust for the current weather type, expressed as: .

[0017] A further improvement of this invention is that the specific process of the adaptive motor control model includes:

[0018] Step 1: Standardize the weather perception feature vector E to [0,1] and then weight and fuse it to obtain the weather influencing factor. wif And extract the control increment from the previous control cycle. ;

[0019] Step 2: Assess driver trust in the current weather type, weather impact factors, vehicle performance changes, and... As the state input of the adaptive motor control model;

[0020] Step 3: Use the negative value of the weighted fusion of the absolute values ​​of the state inputs as the reward function to output the action parameters of the adaptive motor control model. ;

[0021] Step 4: Construct the cost function by combining the output of Step 3;

[0022] Step 5: Solve the cost function using OSQP and output control commands.

[0023] A further improvement of this invention is that the cost function is expressed as:

[0024] ;

[0025] Where N represents the prediction time domain length, Indicates time The predicted first j Step output, Indicates the first j The target output of the step is Dtl, which represents the driver trust level for the current weather type, and Q and R, which represent the weight matrix, determined by the driver-weight adaptation strategy.

[0026] A further improvement of this invention is that the driver-weight adaptation strategy specifically includes determining a weight matrix by introducing driver trust in the current weather type and weather influence factors, calculated as follows:

[0027] ;

[0028] ;

[0029] in, and This represents the baseline weight matrix, which reflects the default penalty intensity under normal weather conditions and a fully trusted state.

[0030] On the other hand, the present invention provides an intelligent control method for a motor controller based on vehicle behavior data, comprising the following steps:

[0031] S1. Collect environmental data, vehicle motion data, and driver status data during vehicle operation;

[0032] S2. Evaluate the vehicle performance changes of drivers and vehicles under different weather conditions, and establish a driver performance evaluation strategy. Evaluate driver proficiency under different weather conditions based on vehicle performance changes, and output driver trust level.

[0033] S3. Construct an adaptive motor control model, allocate weights to the control inputs based on the driver's level of trust, and output control parameters.

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] 1. This invention firstly corrects the control strategy in real time by introducing driver trust level C, which can dynamically match the control intensity according to the proficiency and current state of different drivers, improve the efficiency of human-machine collaboration, and reduce conflicts caused by human intervention.

[0036] 2. By using a weather-weight mapping model and reinforcement learning strategy, it can automatically adapt to various working conditions such as rain, snow, and dryness, without having to redesign the algorithm for each scenario, which greatly simplifies engineering deployment and subsequent maintenance;

[0037] 3. When the driver is not in good condition or the weather is bad, the tracking error penalty is automatically amplified and the smoothing constraint is relaxed, so that the system can quickly and decisively take over or correct the deviation, significantly reducing the risk of skidding, loss of control and collision. Attached Figure Description

[0038] Figure 1 This is a framework diagram of an intelligent control system for a motor controller based on vehicle behavior data according to the present invention;

[0039] Figure 2 This is a flowchart illustrating the construction of an adaptive motor control model for an intelligent control system based on vehicle behavior data, according to the present invention.

[0040] Figure 3 This is a flowchart of an intelligent control method for a motor controller based on vehicle behavior data according to the present invention. Detailed Implementation

[0041] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0042] The term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three cases: A exists alone, A and B exist simultaneously, and B exists alone.

[0043] Example 1

[0044] Figure 1 This embodiment illustrates a framework diagram of an intelligent control system for a motor controller based on vehicle behavior data, including:

[0045] The data acquisition module is used to collect environmental data, vehicle motion data, and driver status data during vehicle operation.

[0046] The weather adaptation model building module is used to evaluate the vehicle performance changes of drivers and vehicles under different weather conditions, establish a driver performance evaluation strategy, evaluate driver proficiency under different weather conditions based on vehicle performance changes, and output driver trust level.

[0047] The adaptive parameter adjustment module is used to construct an adaptive motor control model, which distributes the control inputs with weights according to the driver's trust level and outputs control parameters.

[0048] The data acquisition module includes environmental data, vehicle motion data, and driver status data. The environmental data includes camera images, millimeter-wave radar point clouds, and lidar point clouds. The vehicle motion data includes vehicle speed, acceleration, lateral displacement, wheel speed, and steering angle. The driver status data includes eye tracking, facial expressions, and heart rate monitoring. All data are aligned using timestamps, and noise is eliminated using Kalman filtering.

[0049] The weather adaptation model construction module includes a performance index evaluation unit, a proficiency assessment unit, and a driver trust unit. The performance index evaluation unit is used to obtain a vehicle performance change score based on the absolute value of the difference between the weighted score (normalized for vehicle trajectory error, speed deviation, slippage rate, road friction, and smoothness) and 1. The performance index weights are updated based on a weather-weight mapping model, specifically as follows:

[0050] Construct a weather perception feature vector E, ,in, , representing the normalized rainfall intensity, This represents the normalized snow depth. Indicates the real-time friction coefficient. Indicates visibility. This represents the normalized road surface temperature.

[0051] Based on a small fully connected neural network For historical operating data, the combination of performance index weights that maximizes the change in vehicle performance score is used as the label. The objective function is to minimize the mean squared error (MSE) between the regression output and the label. ,in, It represents the ideal weight vector for simulation or road testing, given for typical rain / snow conditions; it outputs an unnormalized performance index weight sequence vector, and obtains the final performance index weights through Softmax.

[0052] After each evaluation window ends, calculate the vehicle performance change score Cvp for the current window, let the target performance score be Pcvp, and then derive the performance index error. ,when When the error exceeds the preset performance index threshold, perform gradient updates on the weather-weighted mapping model. ,in Indicates the learning rate. Finite difference estimation can be used: fine-tune the weights of each performance metric by a small step and observe the performance changes. The performance metric weight sequence for the k-th iteration, with weights set for each performance metric. ,when The iteration stops when the change value is less than the set threshold.

[0053] This embodiment dynamically adjusts the weights of key indicators such as "road surface temperature," "road friction," and "visibility" for different working conditions such as rain and snow, so that the system pays attention to the risk factors that are most likely to cause accidents at all times, significantly reducing the probability of skidding, loss of control, or collision.

[0054] The proficiency assessment unit incorporates a driver performance evaluation strategy, specifically including: creating a weather type vector W, scoring based on vehicle performance changes, and constructing a long-term proficiency model for the driver under various weather conditions. This is manifested in the following way: for the k-th iteration... , This represents the driver i's proficiency under weather type w, and the initial proficiency of the driver is preset for each weather type. The average driver proficiency level is calculated based on the database and updated in real time. This indicates the performance score for this test.

[0055] The driver trust unit is used to determine driver trust based on the driver's current state and historical state. Specifically, it includes: obtaining the average historical driver state score based on current weather type and historical driver state data, which serves as the driver's historical state baseline (dsh). The driver's historical state is obtained by weighted fusion of eye-tracking indicators, facial expression indicators, and heart rate indicators. Eye-tracking indicators include the eyelid closure ratio (Iprec(t) = number of closed frames(t) / total number of frames(t), with higher values ​​indicating more severe fatigue), and also includes gaze stability. ,in, The standard deviation of the gaze position within the short window is used; a smaller deviation indicates more stable gaze. The eye-tracking index is obtained by weighted summation of the eyelid closure ratio and gaze stability. A pre-trained facial emotion recognition network (such as a ResNet-based micro-expression classification system) is used to output the current "expression stress level." Heart rate belt indicators are expressed as short-term indicators of heart rate variability (HRV). SDNN is the standard deviation of the RR interval; the lower the value, the greater the stress or fatigue.

[0056] The driver's current state score (dsc) within a past time window T with a time interval t is calculated using the same method. The difference between the driver's current state score and the historical state baseline is calculated as the state baseline correction factor (Δds). A negative Δds indicates a poor current state, while a positive Δds indicates an improving state. Allowing the trust score to fluctuate around the baseline yields the driver trust score for the current weather type, denoted as... When the current state is lower than the historical baseline (Δds<0), the confidence level is reduced proportionally; if Δds≥0, no negative reduction is made, and the level is maintained or slightly increased.

[0057] Example 2

[0058] Based on the inventive concept of Embodiment 1, this embodiment proposes a specific construction process for the adaptive motor control model in the adaptive parameter adjustment module. Figure 2 This invention presents a flowchart illustrating the construction of an adaptive motor control model for an intelligent control system based on vehicle behavior data, specifically including:

[0059] Step 1: Standardize the weather perception feature vector E to [0,1] and then weight and fuse it to obtain the weather influencing factor. wif And extract the control increment from the previous control cycle. ;

[0060] Step 2: Combine the current weather type, driver trust level, weather impact factor, vehicle performance change score, and control increment from the previous control cycle. As the state input of the adaptive motor control model;

[0061] Step 3: Use the negative value of the weighted fusion of the absolute values ​​of the state inputs as the reward function to output the action parameters of the adaptive motor control model. ;

[0062] Step three, reinforcement learning, updates the policy network based on the reward function generated after the previous decision, allowing the parameters of the next action to be updated accordingly. It can better balance objectives such as "tracking accuracy, control smoothness, and constraint satisfaction";

[0063] New motion parameters The cost function is then constructed in step four, which makes the model's quadratic objective function focus more on the most important performance dimension at the next moment.

[0064] Step 4: Construct the cost function based on the output of Step 3; the cost function is expressed as:

[0065] ;

[0066] Where N represents the prediction time domain length, indicating that the adaptive motor control model predicts a total of N steps of state and control actions from the current moment into the future. Indicates time The predicted first j Step output, representing measurable quantities of the motor, such as speed or current; DTL indicates the current weather type and driver confidence level. Indicates the first j The target output of the step is given by the upper-level planning (such as vehicle speed cruise), such as the desired speed or torque; Q and R represent the weight matrix, which is determined by the driver-weight adaptation strategy.

[0067] The driver-weighted adaptation strategy specifically involves determining the weight matrix by incorporating driver trust in the current weather type and weather impact factors. The calculation formula is as follows:

[0068] ;

[0069] ;

[0070] in, and This represents the baseline weight matrix, which reflects the default penalty intensity under normal weather conditions and a fully trusted state.

[0071] In the driver trust coefficient, 1 indicates complete trust in the driver, and the system tends to relax tracking penalties and give the driver more room to maneuver; 0 indicates no trust, and the system strengthens tracking control.

[0072] When driver trust decreases, it indicates that the driver is in poor condition or unreliable, the factor (1−Dtl) increases, and Q is amplified accordingly.

[0073] This allows the model to more rigorously reduce tracking errors during optimization, ensuring that the system takes over proactively and suppresses the risk of deviation caused by poor human-machine cooperation.

[0074] The greater the severity of the weather (such as heavy rain or blizzard), the higher the factor. The higher.

[0075] Therefore, the model prioritizes trajectory tracking accuracy in high-risk scenarios such as rain and snow, and resists deviations under adverse conditions such as sideslip and loss of control.

[0076] When driver trust decreases Decreasing causes R to shrink.

[0077] The model calculates the optimal control increment. The inhibition is reduced, allowing for greater and more decisive control actions, and timely correction of deviations.

[0078] In bad weather, Rise, fall This further reduces the penalties for large actions.

[0079] In rain, snow, or low-adhesion scenarios, the system can respond quickly with more aggressive torque or voltage commands, enhancing dynamic stability.

[0080] In situations where safety and comfort are equally important (high trust, good weather), R≈R0 control commands remain smooth, avoiding abrupt acceleration and deceleration.

[0081] In situations of "low trust or high risk," the system automatically "goes free" to achieve the desired tracking as quickly as possible, thereby enhancing security.

[0082] The adaptive motor control model prioritizes minimizing errors when most needed (e.g., when the driver is not in good condition or the weather is bad), while allowing for aggressive control, thereby significantly reducing the risk of accidents. When conditions permit, the system reverts to the default Q0 and R0 settings to avoid excessive intervention and abrupt actions, ensuring a smooth riding experience. The introduction of driver trust allows the control strategy to dynamically match the actual capabilities of different drivers, improving system flexibility and human-machine cooperation. The dual-factor superposition method allows for a smooth transition to various scenarios, achieving versatility of "one framework for multiple operating conditions".

[0083] Step 5: Solve the cost function using OSQP and output control commands; the control commands include voltage vector commands, current / torque reference commands, and speed / position rate commands.

[0084] The threshold and weight settings can be set by default according to the present invention, or they can be set by those skilled in the art.

[0085] Example 3

[0086] Figure 3 This invention presents a flowchart of an intelligent control method for a motor controller based on vehicle behavior data, which is based on the same inventive concept as Embodiments 1 and 2. The intelligent control method for a motor controller based on vehicle behavior data includes the following steps:

[0087] S1. Collect environmental data, vehicle motion data, and driver status data during vehicle operation;

[0088] S2. Evaluate the vehicle performance changes of drivers and vehicles under different weather conditions, and establish a driver performance evaluation strategy. Evaluate driver proficiency under different weather conditions based on vehicle performance changes, and output driver trust level.

[0089] S3. Construct an adaptive motor control model, allocate weights to the control inputs based on the driver's level of trust, and output control parameters.

[0090] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0092] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0093] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0094] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. An intelligent control system for motor control based on vehicle behavior data, characterized in that: include: The data acquisition module is used to collect environmental data, vehicle motion data, and driver status data during vehicle operation. The weather adaptation model building module is used to evaluate the vehicle performance changes of drivers and vehicles under different weather conditions, establish a driver performance evaluation strategy, evaluate driver proficiency under different weather conditions based on vehicle performance changes, and output driver trust level. The adaptive parameter adjustment module is used to construct an adaptive motor control model, which distributes the control inputs with weights according to the driver's trust level and outputs control parameters. The weather adaptation model construction module includes a performance index evaluation unit, a proficiency assessment unit, and a driver trust unit. The performance index evaluation unit is used to obtain a vehicle performance change score based on the absolute value of the difference between the weighted score of the vehicle trajectory error, speed deviation, slip rate, road friction, and smoothness after normalization and 1. The performance index weights are updated based on a weather-weight mapping model, specifically: a weather perception feature vector E is constructed, and based on a small fully connected neural network, the combination of performance index weights that maximizes the vehicle performance change score is searched for as a label on historical operating data. With the goal of minimizing the MSE between the regression output and the label, an unnormalized performance index weight sequence vector is output, and the final performance index weights are obtained through Softmax.

2. The intelligent control system for motor control based on vehicle behavior data according to claim 1, characterized in that: The data acquisition module includes environmental data, vehicle motion data, and driver status data; the environmental data includes camera images, millimeter-wave radar point clouds, and lidar point clouds; the vehicle motion data includes vehicle speed, acceleration, lateral displacement, wheel speed, and steering angle; and the driver status data includes eye tracking, facial expressions, and heart rate monitoring. All data is aligned using timestamps, and noise is eliminated using Kalman filtering.

3. The intelligent control system for motor control based on vehicle behavior data according to claim 2, characterized in that: The performance evaluation unit further includes, after each evaluation window ends, calculating the vehicle performance change score Cvp for the current window, setting the target performance score as Pcvp, and deriving the performance index error. ,when When the error exceeds the preset performance index threshold, perform gradient updates on the weather-weighted mapping model. ,in Indicates the learning rate. This represents the sequence of performance metric weights for the k-th iteration, with each performance metric weight set as follows: ,when The iteration stops when the change value is less than the set threshold.

4. The intelligent control system for motor control based on vehicle behavior data according to claim 1, characterized in that: The proficiency assessment unit incorporates a driver performance evaluation strategy, specifically including: creating a weather type vector W, scoring based on vehicle performance changes, and constructing a long-term proficiency model for the driver under various weather conditions. This is manifested in the following scenario for the k-th iteration: , This represents the driver i's proficiency under weather type w, and the initial proficiency of the driver is preset for each weather type. , This indicates the score for the changes in vehicle performance.

5. The intelligent control system for motor control based on vehicle behavior data according to claim 1, characterized in that: The driver trust unit is used to determine driver trust based on the driver's current state and historical state. Specifically, it includes: obtaining the average historical driver state score based on historical driver state data for the current weather type, using this as the driver's historical state baseline (dsh); and calculating the driver's current state score (dsc) within a past time window T with a time interval of t, using the same method; calculating the difference between the driver's current state score and the historical state baseline as the state baseline correction factor Δds, thus obtaining the driver trust for the current weather type, expressed as: .

6. The intelligent control system for motor control based on vehicle behavior data according to claim 1, characterized in that: The specific process of the adaptive motor control model includes: Step 1: Standardize the weather perception feature vector E to [0,1] and then weight and fuse it to obtain the weather influencing factor. wif And extract the control increment from the previous control cycle. ; Step 2: Assess driver trust in the current weather type, weather impact factors, vehicle performance changes, and... As the state input of the adaptive motor control model; Step 3: Use the negative value of the weighted fusion of the absolute values ​​of the state inputs as the reward function to output the action parameters of the adaptive motor control model. ; Step 4: Construct the cost function by combining the output of Step 3; Step 5: Solve the cost function using OSQP and output control commands.

7. The intelligent control system for motor control based on vehicle behavior data according to claim 6, characterized in that: The cost function is expressed as: ； Where N represents the prediction time domain length, Indicates time The predicted first j Step output, Indicates the first j The target output of the step is Dtl, which represents the driver trust level for the current weather type, and Q and R, which represent the weight matrix, determined by the driver-weight adaptation strategy.

8. The intelligent control system for motor control based on vehicle behavior data according to claim 6, characterized in that: The driver-weight adaptation strategy specifically includes determining a weight matrix by introducing driver trust levels based on the current weather type and weather influencing factors, calculated using the following formula: ； ； in, and represents the baseline weight matrix, which represents the default penalty intensity under normal weather + full trust conditions, and Dtl represents the driver trust level under the current weather type.

9. A motor control method based on vehicle behavior data, used to execute a motor control system based on vehicle behavior data as described in any one of claims 1-8, characterized in that: Includes the following steps: S1. Collect environmental data, vehicle motion data, and driver status data during vehicle operation; S2. Evaluate the vehicle performance changes of drivers and vehicles under different weather conditions, and establish a driver performance evaluation strategy. Evaluate driver proficiency under different weather conditions based on vehicle performance changes, and output driver trust level. S3. Construct an adaptive motor control model, allocate weights to the control inputs based on the driver's level of trust, and output control parameters.

Images