Braking torque distribution method based on artificial intelligence and vehicle dynamics

Abstract

The application relates to a brake torque distribution method based on artificial intelligence and vehicle dynamics. The method obtains current state information of a target vehicle, environment information around the target vehicle, and driving information of a driver on the target vehicle, then inputs the state information, the environment information and the driving information into a target torque distribution model to perform torque distribution, obtains an initial torque distribution result, and finally corrects the initial torque distribution result according to the state information, the driving information and a vehicle dynamics model corresponding to the target vehicle to obtain a target torque distribution result. In the method, the brake torque distribution of the vehicle is performed in combination with the target torque distribution model and the vehicle dynamics model, the dynamic complex working conditions such as sudden change of road adhesion and steering brake coupling can be adapted, and the precise torque distribution effect can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle control, in particular to a brake torque distribution method based on artificial intelligence and vehicle dynamics. BACKGROUND

[0002] With the rapid development of intelligent driving technology, a series of advanced vehicle dynamics control technologies have emerged, among which the electronic stability program (ESP), anti-lock braking system (ABS) and regenerative braking technology are particularly noteworthy. These technologies independently intervene and precisely control the braking force of each wheel to maintain the stability and efficiency of the vehicle during braking.

[0003] At present, the brake torque distribution method of the vehicle mainly distributes the braking force to each wheel through static logic analysis. However, in dynamic complex scenarios such as sudden change of road adhesion and steering braking coupling, the above brake torque distribution method has the problem of inaccurate distribution. SUMMARY

[0004] Therefore, it is necessary to provide a brake torque distribution method based on artificial intelligence and vehicle dynamics that can improve the accuracy of torque distribution.

[0005] In a first aspect, the present application provides a brake torque distribution method based on artificial intelligence and vehicle dynamics, which comprises:

[0006] obtaining the current state information of the target vehicle, the environmental information around the target vehicle, and the driving information of the driver on the target vehicle;

[0007] inputting the state information, the environmental information and the driving information into a target torque distribution model for torque distribution to obtain an initial torque distribution result;

[0008] correcting the initial torque distribution result according to the state information, the driving information and the vehicle dynamics model corresponding to the target vehicle to obtain a target torque distribution result.

[0009] In some embodiments, the initial torque distribution result is corrected according to the state information, the driving information and the vehicle dynamics model corresponding to the target vehicle to obtain the target torque distribution result, which comprises:

[0010] inputting the state information and the driving information into the vehicle dynamics model for dynamics analysis to obtain the maximum braking force and the reference yaw rate of the target vehicle;

[0011] The initial torque distribution result is corrected according to the maximum braking force and the reference yaw rate to obtain a target torque distribution result.

[0012] In some embodiments, the vehicle dynamics model includes a braking force analysis model and an angular velocity analysis model, and the state information and the driving information are input into the vehicle dynamics model for dynamics analysis to obtain the maximum braking force and the reference yaw rate of the target vehicle, including:

[0013] The state information is input into the braking force analysis model for braking force analysis to obtain the maximum braking force of the target vehicle.

[0014] The state information and the driving information are input into the angular velocity analysis model for analysis to obtain the reference yaw rate of the target vehicle.

[0015] In some embodiments, the initial torque distribution result includes an initial braking torque, and the target torque distribution result includes a target braking torque, and the initial torque distribution result is corrected according to the maximum braking force and the reference yaw rate to obtain the target torque distribution result, including:

[0016] It is determined whether the initial torque distribution result satisfies a preset braking force condition and / or a preset angular velocity condition; the preset braking force condition includes that a tire braking force corresponding to the initial braking torque does not exceed the maximum braking force; and the preset angular velocity condition includes that an angular velocity difference between a yaw rate corresponding to the initial braking torque and the reference yaw rate does not exceed a preset threshold value.

[0017] If the initial torque distribution result satisfies the preset braking force condition and the preset angular velocity condition, the initial braking torque is taken as the target braking torque.

[0018] If the initial torque distribution result does not satisfy the preset braking force condition or the preset angular velocity condition, the initial braking torque is adjusted to obtain an adjusted braking torque, and the step of determining whether the initial torque distribution result satisfies the preset braking force condition and / or the preset angular velocity condition is executed again according to the adjusted braking torque.

[0019] In some embodiments, the initial braking torque is adjusted to obtain the adjusted braking torque, including:

[0020] If the initial torque distribution result does not satisfy the preset braking force condition, a smaller value between a braking torque corresponding to the maximum braking force and the initial braking torque is taken as the target braking torque.

[0021] If the initial torque distribution result does not satisfy the angular velocity condition, a torque increment is determined according to the angular velocity difference, and the initial braking torque is adjusted according to the torque increment to obtain the adjusted braking torque.

[0022] In some embodiments, the method further comprises:

[0023] constructing a simulation scenario library of the test vehicle; the simulation scenario library comprises basic regulation working conditions, extreme dangerous working conditions and new energy working conditions;

[0024] controlling the test vehicle to perform driving tests under various working conditions in the simulation scenario library, collecting dynamic data of the test vehicle under each working condition to form a sample data set;

[0025] training the initial torque distribution model according to the sample data set to obtain a target torque distribution model.

[0026] In some embodiments, the sample data set includes state sample data, action sample data and state transition sample data, the initial torque distribution model includes an initial decision network and an initial state prediction network, and the initial torque distribution model is trained according to the sample data set to obtain a target torque distribution model, including:

[0027] inputting the state sample data into the initial decision network to make a braking decision, and outputting decision action information;

[0028] inputting the decision action information and the state sample data into the initial state prediction network to perform state prediction, and obtaining state transition information;

[0029] determining a target loss according to the decision action information, the state transition information and the sample data set;

[0030] adjusting parameters in the initial decision network according to the target loss until the target loss meets a preset error threshold requirement, and taking the trained initial decision network as the target torque distribution model.

[0031] In a second aspect, the present application also provides a braking torque distribution device based on artificial intelligence and vehicle dynamics, which comprises:

[0032] an acquisition module configured to acquire current state information of a target vehicle, environmental information around the target vehicle, and driving information of a driver on the target vehicle;

[0033] a distribution module configured to input the state information, the environmental information and the driving information into a target torque distribution model to perform torque distribution, and obtain an initial torque distribution result;

[0034] a correction module configured to correct the initial torque distribution result according to the state information, the driving information and a vehicle dynamics model corresponding to the target vehicle, and obtain a target torque distribution result.

[0035] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0036] obtaining current state information of the target vehicle, environment information around the target vehicle, and driving information of a driver on the target vehicle;

[0037] inputting the state information, the environment information, and the driving information into a target torque distribution model to perform torque distribution, to obtain an initial torque distribution result;

[0038] correcting the initial torque distribution result according to the state information, the driving information, and a vehicle dynamics model corresponding to the target vehicle, to obtain a target torque distribution result.

[0039] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements the following steps when executed by a processor:

[0040] obtaining current state information of the target vehicle, environment information around the target vehicle, and driving information of a driver on the target vehicle;

[0041] inputting the state information, the environment information, and the driving information into a target torque distribution model to perform torque distribution, to obtain an initial torque distribution result;

[0042] correcting the initial torque distribution result according to the state information, the driving information, and a vehicle dynamics model corresponding to the target vehicle, to obtain a target torque distribution result.

[0043] In a fifth aspect, the present application provides a computer program product, comprising a computer program, and the computer program implements the following steps when executed by a processor:

[0044] obtaining current state information of the target vehicle, environment information around the target vehicle, and driving information of a driver on the target vehicle;

[0045] inputting the state information, the environment information, and the driving information into a target torque distribution model to perform torque distribution, to obtain an initial torque distribution result;

[0046] correcting the initial torque distribution result according to the state information, the driving information, and a vehicle dynamics model corresponding to the target vehicle, to obtain a target torque distribution result.

[0047] The above-mentioned braking torque distribution method based on artificial intelligence and vehicle dynamics, the method obtains the current state information of the target vehicle, the environmental information around the target vehicle, and the driving information of the driver on the target vehicle, then inputs the state information, environmental information and driving information into the target torque distribution model for torque distribution, obtains the initial torque distribution result, and finally corrects the initial torque distribution result according to the state information, driving information and vehicle dynamics model corresponding to the target vehicle, to obtain the target torque distribution result. In the above method, the target torque distribution model accurately responds to the change of road adhesion and the complex driving scene, and the vehicle dynamics model corrects the initial torque distribution result, which can make the braking process more stable, can adapt to the dynamic complex working conditions such as sudden change of road adhesion and steering braking coupling, avoid the risk of low adhesion side wheel lock and vehicle instability, and then realize precise torque distribution effect. In addition, the above method combines the generalization decision ability of artificial intelligence and the physical constraint mechanism of vehicle dynamics, which can adaptively adjust the braking torque distribution and improve the comprehensive performance and user experience of the vehicle. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 It is an internal structure diagram of a computer device in some embodiments;

[0049] Figure 2 It is one of the flowchart diagrams of the braking torque distribution method based on artificial intelligence and vehicle dynamics;

[0050] Figure 3 It is the second flowchart diagram of the braking torque distribution method based on artificial intelligence and vehicle dynamics;

[0051] Figure 4 It is the third flowchart diagram of the braking torque distribution method based on artificial intelligence and vehicle dynamics;

[0052] Figure 5 It is the fourth flowchart diagram of the braking torque distribution method based on artificial intelligence and vehicle dynamics;

[0053] Figure 6 It is the fifth flowchart diagram of the braking torque distribution method based on artificial intelligence and vehicle dynamics;

[0054] Figure 7 It is the sixth flowchart diagram of the braking torque distribution method based on artificial intelligence and vehicle dynamics;

[0055] Figure 8 It is a structure block diagram of the braking torque distribution device based on artificial intelligence and vehicle dynamics. DETAILED DESCRIPTION

[0056] The term "and / or" in the embodiments of the present application describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can represent the following three cases: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after it.

[0057] The term "multiple" in the embodiments of the present application means two or more, and other quantifiers are similar.

[0058] The term "at least one" in the embodiments of the present application means one or more, for example, at least one of A, B and C can mean the following six cases: A exists alone, B exists alone, C exists alone, A and B exist simultaneously, A and C exist simultaneously, B and C exist simultaneously, and A, B and C exist simultaneously.

[0059] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0060] With the rapid development of intelligent driving technology, a series of advanced vehicle dynamics control technologies have emerged, among which electronic stability program (ESP), anti-lock braking system (ABS) and regenerative braking technology are particularly noteworthy. These technologies independently intervene and accurately control the braking force of each wheel to maintain the stability and efficiency of the vehicle during braking. The existing vehicle braking torque distribution method is difficult to adapt to dynamic complex conditions such as sudden change of road adhesion and steering braking coupling because it relies on static rules and accurate dynamic models, and cannot cooperatively optimize the conflict of multiple targets such as braking distance, lateral stability, energy recovery and comfort, resulting in a sharp increase in the risk of low adhesion side wheel lock and vehicle instability, and a sharp decline in performance under parameter disturbance. Therefore, the above-mentioned braking torque distribution method has the problem of inaccurate distribution.

[0061] In view of this, the embodiments of the present application propose a braking torque distribution method based on artificial intelligence and vehicle dynamics, which can integrate the generalization decision-making ability of artificial intelligence and the physical constraint mechanism of vehicle dynamics, realize multi-target adaptive distribution without relying on high-precision models, improve the accuracy of vehicle torque distribution, and significantly improve the safety and comprehensive performance of the vehicle under complex conditions.

[0062] It should be noted that the beneficial effects brought by the embodiments of the present application or the technical problems solved are not limited to the above, but also other implicit or related problems, which can be seen from the following description of the embodiments.

[0063] The technical solutions of the present application and how the technical solutions solve the above technical problems will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described in detail in some embodiments. The embodiments of the present application will be described below with reference to the accompanying drawings.

[0064] In some embodiments, the brake torque distribution method based on artificial intelligence and vehicle dynamics provided by the embodiments of the present application can be applied to a computer device as shown in Figure 1 The internal structure diagram of the computer device can be as shown in Figure 1 The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be achieved through WIFI, mobile cellular network, NFC (Near Field Communication) or other technologies. The computer program is executed by the processor to implement a brake torque distribution method based on artificial intelligence and vehicle dynamics. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0065] Those skilled in the art can understand that Figure 1 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0066] In some embodiments, as shown in Figure 2 An artificial intelligence and vehicle dynamics based brake torque distribution method is provided, which is applied to Figure 1 a computer device as an example, including the following steps:

[0067] S201, obtaining the current state information of the target vehicle, the surrounding environment information of the target vehicle, and the driving information of the driver on the target vehicle.

[0068] The state information includes at least one of the longitudinal vehicle speed , the yaw rate , the longitudinal acceleration , the lateral acceleration , the maximum friction coefficient between each wheel and the road surface (including , , , ), the wheel slip rate of each wheel (including , , , ), etc. Among them, represents the left front wheel, represents the right front wheel, represents the left rear wheel, represents the right rear wheel.

[0069] The environment information includes at least one of the traffic participants (pedestrians, other vehicles, obstacles, etc.), the road type (highway, urban road), the traffic rules (traffic lights, speed limit signs, road markings, etc.), etc.

[0070] The driving information includes at least one of the steering wheel angle , the brake pedal opening , the pedal stroke (acceleration / brake pedal), the steering force, the driving intention (such as sudden acceleration, light braking, lane change), etc.

[0071] In the embodiments of the present application, the computer device can obtain the current state information of the target vehicle, the environmental information around the target vehicle, and the driving information of the driver on the target vehicle through various types of sensors arranged on the target vehicle. Specifically, the current state information of the target vehicle and the environmental information around the target vehicle can be directly obtained by using sensors. Alternatively, the environmental information around the target vehicle and the running data (such as vehicle speed, wheel speed) of the vehicle itself can be collected by using millimeter wave radar, camera, laser radar, etc., and then the vehicle state estimation is performed according to the environmental information and the running data of the vehicle itself to obtain the current state information of the target vehicle. The signals such as steering wheel angle, pedal stroke (acceleration / braking pedal), steering force, etc. can be obtained through the sensors on the target vehicle to identify the driving intention of the driver (such as sudden acceleration, light braking, lane change). Alternatively, after the computer device obtains the data output by the sensors, the data output by the sensors can be subjected to preprocessing operations such as filtering and noise reduction (removing sensor errors and interference), time and space synchronization (unifying the time and space reference of different sensor data), feature extraction (extracting key parameters such as obstacle relative speed and vehicle slip rate) to obtain the current state information of the target vehicle, the environmental information around the target vehicle, and the driving information of the driver on the target vehicle.

[0072] S202, inputting the state information, the environmental information and the driving information into a target torque distribution model to perform torque distribution, and obtaining an initial torque distribution result.

[0073] The target torque distribution model can be a neural network model or a machine learning model. The initial torque distribution result includes an initial braking torque of each wheel of the target vehicle.

[0074] In the embodiments of the present application, the computer device can pre-train the initial torque distribution model according to a plurality of different driving conditions to train a model capable of accurately distributing torque to each tire of the vehicle, i.e. to obtain the target torque distribution model. After the computer device obtains the current state information of the target vehicle, the environmental information around the target vehicle, and the driving information of the driver on the target vehicle based on the above steps, the computer device can input the state information, the environmental information and the driving information into the target torque distribution model to distribute torque to each wheel of the target vehicle through the target torque distribution model, and obtain the initial torque distribution result.

[0075] S203, correcting the initial torque distribution result according to the state information, the driving information and a vehicle dynamics model corresponding to the target vehicle to obtain a target torque distribution result.

[0076] The target torque distribution result includes a target braking torque of each wheel of the target vehicle.

[0077] In the embodiments of the present application, the computer device can pre-construct vehicle dynamics models corresponding to different types of vehicles. After obtaining the current state information of the target vehicle and the driving information of the driver on the target vehicle, the computer device can first determine the vehicle dynamics model corresponding to the target vehicle, and then input the state information and the driving information into the vehicle dynamics model for dynamics analysis, analyze the vehicle dynamics limit information of the target vehicle in the normal scene, the extreme, complex or high-risk scene, such as the vehicle dynamics limit information in the “locked wheel” and “insufficient braking force” scene, and then correct the initial torque distribution result based on the vehicle dynamics limit information to obtain the target torque distribution result, so as to ensure that the target vehicle will not exceed the vehicle dynamics limit during the vehicle driving process according to the target torque distribution result, so as to balance the braking efficiency and the vehicle stability (prevent side slip and spin). The vehicle dynamics limit information can include at least one of the maximum braking torque, the optimal braking deceleration, the braking distance limit, the maximum lateral acceleration, the tire friction limit, and the maximum yaw rate of the target vehicle.

[0078] The method for distributing braking torque based on artificial intelligence and vehicle dynamics provided in the embodiments of the present application can obtain the current state information of the target vehicle, the environmental information around the target vehicle, and the driving information of the driver on the target vehicle, input the state information, the environmental information and the driving information into the target torque distribution model for torque distribution to obtain the initial torque distribution result, and finally correct the initial torque distribution result according to the state information, the driving information and the vehicle dynamics model corresponding to the target vehicle to obtain the target torque distribution result. In the above method, the target torque distribution model accurately responds to the road adhesion change and the complex driving scene, and the initial torque distribution result is corrected by the vehicle dynamics model, so that the braking process is more stable, can adapt to road adhesion mutation and steering-braking coupling and other dynamic complex working conditions, avoids the risk of low adhesion side wheel locking and vehicle instability, and thus can achieve precise torque distribution effect. In addition, the above method combines the generalization decision-making ability of artificial intelligence and the physical constraint mechanism of vehicle dynamics, can adaptively adjust the braking torque distribution, and improves the comprehensive performance and user experience of the vehicle.

[0079] In some embodiments, a specific implementation of correcting the initial torque distribution result is also provided, as shown in Figure 3 The “correcting the initial torque distribution result according to the state information, the driving information and the vehicle dynamics model corresponding to the target vehicle to obtain the target torque distribution result” in S203 includes:

[0080] S301, inputting the state information and the driving information into the vehicle dynamics model for dynamics analysis to obtain the maximum braking torque and the reference yaw rate of the target vehicle.

[0081] The reference yaw rate can be the maximum yaw rate or the optimal yaw rate under the current driving scenario.

[0082] In this embodiment of the application, after obtaining the current state information of the target vehicle and the driving information of the driver on the target vehicle, the computer device can first determine the vehicle dynamics model corresponding to the target vehicle, and then input the state information and driving information into the vehicle dynamics model for dynamic analysis, and analyze the maximum braking force and reference yaw rate of the target vehicle during the driving process.

[0083] Specifically, the aforementioned vehicle dynamics model includes a braking force analysis model and an angular velocity analysis model. Based on this, such as Figure 4 As shown, S301 above includes:

[0084] S3011, input the status information into the braking force analysis model to perform braking force analysis and obtain the maximum braking force of the target vehicle.

[0085] In this embodiment, the computer device can input state information into the braking force analysis model to perform braking force analysis and obtain the maximum braking force of the target vehicle. Specifically, the braking force analysis model can first analyze the vertical force of each wheel in the target vehicle based on the state information of the target vehicle, and then, based on the vertical force of each wheel and the maximum friction coefficient between the tire and the road surface of each wheel (…),… , , , Multiple longitudinal force values ​​are obtained, and the largest longitudinal force value is taken as the maximum braking force. For example, the longitudinal force is the product of the friction coefficient and the vertical force. The braking force analysis model can be expressed by the following relationship:

[0086]

[0087] in, The left front wheel of the target vehicle Vertical force; The right front wheel of the target vehicle Vertical force; The left rear wheel of the target vehicle Vertical force; The right rear wheel of the target vehicle The vertical force; m is the total mass of the vehicle; a is the distance from the center of gravity to the front axle; b is the distance from the center of gravity to the rear axle; L represents the wheelbase, L=a+b; h is the height of the center of gravity; It is longitudinal acceleration; It is lateral acceleration; This refers to the track width of the front wheels; Track for rear wheels.

[0088] S3012, input the state information and the driving information into the angular velocity analysis model for analysis to obtain the reference yaw rate of the target vehicle.

[0089] In the embodiment of the application, the computer device can input the state information and the driving information into the angular velocity analysis model, analyze the yaw rate of the target vehicle through the angular velocity analysis model, and obtain the reference yaw rate of the target vehicle. The angular velocity analysis model can be represented by the following relationship:

[0090]

[0091] wherein, is the reference yaw rate; is the longitudinal vehicle speed; is the understeering coefficient, which is a preset value 0.002 ; is the steering wheel angle;

[0092] S302, correcting the initial torque distribution result according to the maximum braking force and the reference yaw rate to obtain a target torque distribution result.

[0093] In the embodiment of the application, after the computer device obtains the maximum braking force, the reference yaw rate and the initial torque distribution result, the initial torque distribution result can be analyzed to analyze the braking force and the yaw rate corresponding to the initial braking torque, and then the size relationship between the braking force and the maximum braking force is compared, and the size relationship between the yaw rate and the reference yaw rate is compared to obtain a comparison result. Then, the initial torque distribution result is corrected according to the comparison result to obtain a target torque distribution result.

[0094] Specifically, the initial torque distribution result includes an initial braking torque, and the target torque distribution result includes a target braking torque. On this basis, as shown in Figure 5 S302 includes:

[0095] S3021, determining whether the initial torque distribution result meets a preset braking force condition and / or a preset angular velocity condition.

[0096] The preset braking force condition includes that the tire braking force corresponding to the initial braking torque does not exceed the maximum braking force; and the preset angular velocity condition includes that the angular velocity difference between the yaw rate corresponding to the initial braking torque and the reference yaw rate does not exceed a preset threshold.

[0097] In the embodiment of the present application, after obtaining the initial braking torque in the initial torque distribution result, the computer device determines the tire braking force according to the quotient of the initial braking torque and the tire radius, and then calculates the yaw rate through the two-degree-of-freedom vehicle dynamics model. The computer device can compare the tire braking force with the maximum braking force to determine whether the initial torque distribution result meets the preset braking force condition, and compare the yaw rate with the reference yaw rate to determine whether the initial torque distribution result meets the preset angular velocity condition. If the tire braking force corresponding to the initial braking torque is greater than the maximum braking force, it is determined that the initial torque distribution result does not meet the preset braking force condition, and if the tire braking force corresponding to the initial braking torque is less than or equal to the maximum braking force, it is determined that the initial torque distribution result meets the preset braking force condition. The yaw rate and the reference yaw rate are subtracted to obtain the angular velocity difference between the yaw rate and the reference yaw rate. For example, if the angular velocity difference is greater than a preset threshold, it is determined that the initial torque distribution result does not meet the preset angular velocity condition, and if the angular velocity difference is less than or equal to the preset threshold, it is determined that the initial torque distribution result meets the preset angular velocity condition.

[0098] S3022, if the initial torque distribution result meets the preset braking force condition and the preset angular velocity condition, the initial braking torque is taken as the target braking torque.

[0099] In the embodiment of the present application, if the initial torque distribution result meets the preset braking force condition and the preset angular velocity condition, the initial braking torque is taken as the target braking torque.

[0100] S3023, if the initial torque distribution result does not meet the preset braking force condition or the preset angular velocity condition, the initial braking torque is adjusted to obtain an adjusted braking torque, and the step of determining whether the initial torque distribution result meets the preset braking force condition and / or the preset angular velocity condition is executed again according to the adjusted braking torque.

[0101] In the embodiment of the present application, if the initial torque distribution result does not meet the preset braking force condition or the preset angular velocity condition, the initial braking torque is adjusted upward or downward to obtain an adjusted braking torque, and the step of determining whether the initial torque distribution result meets the preset braking force condition and / or the preset angular velocity condition is executed again according to the adjusted braking torque, that is, the step of S3021 is executed again. If the initial torque distribution result does not meet the preset braking force condition, the smaller value between the maximum braking force corresponding braking torque and the initial braking torque is taken as the target braking torque, that is, , is the target braking torque, is the initial braking torque, This represents the braking torque corresponding to the maximum braking force. If the initial torque distribution does not meet the angular velocity condition, the torque increment is determined based on the angular velocity difference, and the initial braking torque is adjusted upwards or downwards based on the torque increment to obtain the adjusted braking torque. This process of determining the torque increment based on the angular velocity difference can be expressed by the following relationship:

[0102]

[0103] in, For torque increment, The difference in angular velocity; This is the proportionality coefficient.

[0104] In some embodiments, such as Figure 6 As shown, the above-mentioned braking torque distribution method based on artificial intelligence and vehicle dynamics also includes:

[0105] S401, Construct a simulation scenario library for the test vehicle.

[0106] The simulation scenario library includes basic regulatory operating conditions, extreme dangerous operating conditions, and new energy operating conditions.

[0107] In this embodiment, to simulate a real operating environment and cover diverse application scenarios, a simulation scenario library is first constructed. This scenario library integrates typical operating conditions of the target application scenarios, including but not limited to different environmental conditions (such as road conditions and weather), task requirements (such as braking, cruising, and obstacle avoidance), and interference factors (such as sensor noise and communication latency), providing a diverse and highly realistic virtual testing environment for subsequent training. The process of constructing various operating conditions in the simulation scenario library is as follows:

[0108] (1) Construct basic regulatory operating conditions to cover safety standard requirements, specifically including high-adhesion straight-line braking condition, low-adhesion braking condition, and steering-braking composite condition. That is: ① High-adhesion straight-line braking condition: uniform road surface adhesion coefficient initial velocity Braking strength is ②Low-adhesion braking condition: uniform low-adhesion road surface , Verify anti-lock braking capability. ③ Steering and braking combined condition: steering wheel angle (corresponding radius of curvature) ), and apply braking intensity simultaneously. The braking force was measured to verify the stability control function.

[0109] (2) Constructing extreme dangerous operating conditions to enhance adaptability to edge scenarios, specifically including split-road braking conditions, abrupt change braking conditions on connected road surfaces, and high-speed obstacle avoidance conditions. That is: ① Split-road braking conditions: left side road surface right side ; initial speed , emergency braking . ② Abrupt braking condition of docking pavement: high adhesion (μf=0.8) in the front section and low adhesion (μr=0.2) in the rear section; abrupt point vehicle speed v0=60 km / h. . ③ High-speed obstacle avoidance condition: double lane shift test (ISO 3888-2), vehicle speed v0=80 km / h, synchronous full braking. (3) New energy working conditions are constructed, specifically including regenerative braking coordination working condition and energy recovery priority working condition. That is: ① Regenerative braking coordination working condition: motor regenerative braking proportion is 50% under medium-low speed (v0=40 km / h); composite motor torque response delay is 0.1 s. ② Energy recovery priority working condition: long downhill road section (slope range -10°<θ<10°), maintain constant speed by regenerative braking.

[0110] (3) New energy working conditions are constructed, specifically including regenerative braking coordination working condition and energy recovery priority working condition. That is: ① Regenerative braking coordination working condition: motor regenerative braking proportion is 50% under medium-low speed (v0=40 km / h); composite motor torque response delay is 0.1 s. ② Energy recovery priority working condition: long downhill road section (slope range -10°<θ<10°), maintain constant speed by regenerative braking.

[0111] S402, control the test vehicle to drive in various working conditions in the simulation scenario library, collect the dynamics data of the test vehicle under each working condition, and form a sample data set.

[0112] The sample data set D includes state sample data Ds, action sample data Da, and state transition sample data Ds.

[0113] In the embodiments of the present application, based on the constructed simulation scenario library, the simulation environment is driven to run, and the test vehicle is controlled to drive in various working conditions in the simulation scenario library. In the process of running the simulation environment, the test vehicle collects dynamics data under each working condition, such as motion state parameters (such as speed, acceleration, pose), environmental interaction parameters (such as braking force, driving force, force feedback), etc., finally cleans and labels the collected dynamics data, removes incorrect or invalid data, and adds corresponding labels to the data, to form a sample data set D={( , , )} for training.

[0114] Ds is the state sample data, that is, the data obtained by the sensor on the test vehicle, including longitudinal vehicle speed, yaw rate, longitudinal acceleration, lateral acceleration, steering wheel angle, brake pedal opening, road adhesion coefficient under four wheels, and wheel slip ratio, etc. Da is the action sample data, that is, the four-wheel target braking torque vector, at a certain time step t in the simulation, according to the current state of the test vehicle ​​​​​​​​​​an outputted four-dimensional vector, [ , , , ] is a target braking torque value allocated by the network for the four wheels. is state transition sample data, and the specific process is as follows: at time step t, the environment is in state , the agent performs action , that is, four braking torque instructions are applied to the simulation test vehicle model, the simulation dynamics model calculates a series of dynamic responses according to physical laws (Newtonian mechanics, tire magic formula, etc.) based on the current vehicle state and the input torque instruction, and after a simulation step, the environment is updated to a new state .

[0115] S403, training the initial torque allocation model according to the sample data set to obtain a target torque allocation model.

[0116] wherein the initial torque allocation model is a deep reinforcement learning network.

[0117] In the embodiments of the application, after obtaining the sample data set, the computer device can input the sample data set to the initial torque allocation model, adjust the parameters in the initial torque allocation model according to the error between the output result of the initial torque allocation model and the standard result, and obtain an adjusted initial torque allocation model, that is, a target torque allocation model.

[0118] Specifically, the above sample data set includes state sample data, action sample data, and state transition sample data, and on this basis, as shown in Figure 7 , the above S403 includes:

[0119] S4031, inputting the state sample data into the initial decision network to make a braking decision, and outputting decision action information.

[0120] S4032, inputting the decision action information and the state sample data into the initial state prediction network to make a state prediction, and obtaining state transition information.

[0121] S4033, determining a target loss according to the decision action information, the state transition information, and the sample data set.

[0122] S4034, adjusting the parameters in the initial decision network according to the target loss until the target loss meets the preset error threshold requirement, and taking the trained initial decision network as a target torque allocation model.

[0123] wherein the initial decision network is a neural network. The decision action information corresponds to the action sample data. The state transition information corresponds to the state transition sample data.

[0124] In the embodiments of the present application, the computer device can obtain a sample data set, and then input the state sample data into the initial decision network to make a braking decision and output decision action information. After obtaining the decision action information, the computer device can input the decision action information and the state sample data into the initial state prediction network to perform state prediction and obtain state transition information. The computer device can determine a target loss based on the decision action information, the state transition information and the sample data set. Specifically, the decision loss can be determined based on the decision action information, the action sample data, the state transition information and the state transition sample data, and the penalty loss can be determined based on the decision action information and the state transition information, and finally the decision loss and the penalty loss are summed to obtain the target loss. After determining the target loss, the computer device can adjust the parameters in the initial decision network according to the target loss until the target loss meets the preset error threshold requirement, and the trained initial decision network is used as the target torque distribution model.

[0125] For example, the model training specifically includes the following steps:

[0126] (1) Design a reward function R (i.e., decision loss) to quantify the pros and cons of the agent's actions in the simulation scenario. The reward function integrates multiple dimensions, such as safety (avoiding collisions, deviations), efficiency (energy consumption, time cost), comfort (acceleration fluctuation, impact degree), etc. By setting reasonable reward rules (positively encouraging optimal strategies, negatively punishing rule violations / inefficient behaviors), the deep reinforcement learning model learns the "optimal decision mode". Through reward function feedback, the policy network parameters are iteratively updated to optimize the decision logic. At the same time, an optional online adaptive mechanism is supported. During training or after deployment, the strategy can be dynamically adjusted according to real-time data (such as environmental changes and model drift), enhancing the model's generalization and robustness.

[0127]

[0128] wherein the weight coefficient According to the target priority setting, is the output of the dynamics reference model; represents the speed drop within 0.1s, used to maximize deceleration; represents the yaw rate deviation, used to minimize the yaw rate deviation; is the slip rate, used to suppress excessive slip ( penalized when >0.2); is the regenerative braking energy recovery, used to maximize energy recovery; is the vehicle body pitch angle, used to improve comfort. Represents braking efficiency, the greater the weight, the model tends to shorten the braking distance, encourage to produce greater longitudinal deceleration; Represents lateral stability, the greater the weight, the model tends to suppress yaw rate error, i.e. prevent the vehicle from fishtailing or pushing, maintain a stable driving trajectory; Represents tire adhesion safety, the greater the weight, the model punishes more severely for slip ratio exceeding the safety threshold (0.2), forcing the model to control wheel slip rate near the adhesion peak, avoiding lockup leading to loss of control; Represents energy recovery efficiency, the greater the weight, the model tends to allocate more braking torque to the drive motor to maximize the recovery of kinetic energy and improve the range of electric vehicles. Represents ride comfort, the greater the weight, the model tends to reduce the pitch motion of the vehicle body, i.e. suppress the "nodding" phenomenon when braking, improving the driving comfort.

[0129] Multi-objective optimization strategy is an intelligent decision-making framework that balances multiple interdependent objectives through a systematic approach in complex decision-making scenarios. Based on mathematical modeling, it converts multiple requirements in practical problems (such as efficiency, cost, safety, stability, etc.) into quantifiable optimization objectives, and combines constraint conditions (such as resource limitations, physical limits, rule requirements) to build a multi-objective optimization model. Dynamic weight distribution mechanism adjusts target priority in real time according to driving mode, avoiding suboptimal solutions caused by fixed weights. Define the weight vector:

[0130]

[0131] Where the priority factor is calculated in real time based on the current state of the vehicle (such as speed, braking intensity, yaw error, driving mode flag). For example, the emergency flag will directly increase the priority of (braking distance). The sensitivity coefficient controls the rate of change of the control weight ( , calibrated through simulation). The priority factor is designed as follows:

[0132]

[0133] (2) Add a boundary penalty term (i.e. penalty loss) to the loss function:

[0134]

[0135] (3) Update the policy network parameters :

[0136]

[0137] where, Represents the average value of a batch of empirical data (a set of time steps t). A ratio, representing the probability of the new policy (to be updated) performing an action Divided by the probability of the old policy (before updating) performing the same action. Represents the generalized advantage function (GAE), the discount factor γ = 0.99; clip is the clipping ratio to ensure training stability ( ).

[0138] (4) Performance verification and iteration: During the training process, a performance verification link is set. According to the preset evaluation index (such as decision accuracy, response speed, robustness), the model performance is periodically detected. If it does not meet the requirements, return to the strategy training link, adjust the training parameters (such as learning rate, exploration rate) or expand the scene data, and reiterate optimization; if the performance meets the requirements, enter the next stage. When the model performance passes the verification, the model is frozen, the network parameters and structure are fixed, and a standardized and deployable intelligent decision-making model is formed. This model can be integrated into actual systems (such as intelligent driving controllers, industrial automation equipment), realizing the transformation from virtual training to physical application, supporting the stable and efficient execution of decision-making tasks by agents in real scenarios.

[0139] (5) Key scene test results, including two scenarios:

[0140] Scenario 1: Low adhesion on open road emergency braking ( , , )

[0141]

[0142] Scenario 2: Variable adhesion curve braking ( , R = 100m, )

[0143]

[0144] Based on the above all embodiments, a brake torque distribution method based on artificial intelligence and vehicle dynamics is also provided, which comprises:

[0145] S501, a simulation scene library of the test vehicle is constructed. The simulation scene library includes basic regulation working conditions, extreme danger working conditions and new energy working conditions.

[0146] S502, control the test vehicle to drive in various working conditions in the simulation scene library, collect the dynamics data of the test vehicle in each working condition, and form a sample data set.

[0147] S503, input the state sample data into the initial decision network to make a braking decision, output decision action information, and input the decision action information and the state sample data into the initial state prediction network to make a state prediction, and obtain state transition information.

[0148] S504, determine a target loss according to the decision action information, the state transition information, and the sample data set, adjust parameters in the initial decision network according to the target loss until the target loss meets a preset error threshold requirement, and take the trained initial decision network as a target torque distribution model.

[0149] S505, obtain current state information of the target vehicle, environment information around the target vehicle, and driving information of a driver on the target vehicle.

[0150] S506, input the state information, the environment information, and the driving information into the target torque distribution model to make torque distribution, obtain an initial torque distribution result, and input the state information into the braking force analysis model to make braking force analysis, and obtain a maximum braking force and a reference yaw rate of the target vehicle.

[0151] S507, determine whether the initial torque distribution result meets a preset braking force condition and / or a preset angular velocity condition. The preset braking force condition includes that a tire braking force corresponding to the initial braking torque does not exceed the maximum braking force. The preset angular velocity condition includes that an angular velocity difference between a yaw rate corresponding to the initial braking torque and the reference yaw rate does not exceed a preset threshold.

[0152] S508, when the initial torque distribution result meets the preset braking force condition and the preset angular velocity condition, take the initial braking torque as a target braking torque.

[0153] S509, when the initial torque distribution result does not meet the preset braking force condition or the preset angular velocity condition, if the initial torque distribution result does not meet the preset braking force condition, take a smaller value between a braking torque corresponding to the maximum braking force and the initial braking torque as the target braking torque; if the initial torque distribution result does not meet the angular velocity condition, determine a torque increment according to the angular velocity difference, adjust the initial braking torque according to the torque increment to obtain an adjusted braking torque, and return to execute the step S507 according to the adjusted braking torque.

[0154] The method described in the embodiments of the application can realize dynamic working condition adaptation by real-time decision of the target torque distribution model instead of static rules, responding to road adhesion changes and complex driving scenarios. By fusing vehicle dynamics principles and data-driven optimization, braking efficiency, stability and energy recovery can be synergistically improved, and the multi-objective conflict bottleneck can be broken through. By combining a physical constraint verification mechanism, the dependence on accurate models is eliminated, and reliable control under parameter perturbation is guaranteed, which can build a strong robust control architecture. In the dynamics constraint module and feedback verification, based on the vehicle dynamics model (considering tire-road friction, suspension response, and vehicle mass distribution), the "braking demand" of the target torque distribution model decision is converted into an executable brake force distribution scheme (such as front and rear wheel brake force ratio and electronic stability program intervention intensity), taking into account braking efficiency and vehicle stability (preventing sideslip and spin), verifying the braking strategy output by the target torque distribution model in combination with vehicle dynamics limits (such as tire adhesion and brake force distribution) and real-time state (current vehicle speed and road friction coefficient), avoiding decisions that exceed the physical capabilities of the vehicle (such as "locked wheels" and "insufficient brake force"), and returning to adjust the decision if the constraints are not met, ensuring safety and feasibility. The application overcomes the above technical bottlenecks by fusing the generalization decision-making ability of artificial intelligence and the physical constraint mechanism of vehicle dynamics, realizes multi-objective adaptive distribution without relying on high-precision models, and significantly improves vehicle safety and comprehensive performance in complex working conditions.

[0155] The method described in each of the above steps is described in the foregoing embodiments, and the details are described in the foregoing description, which will not be repeated here.

[0156] It should be understood that although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0157] Based on the same inventive concept, the embodiments of the present application also provide an artificial intelligence and vehicle dynamics based brake torque distribution device for implementing the artificial intelligence and vehicle dynamics based brake torque distribution method described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more artificial intelligence and vehicle dynamics based brake torque distribution device embodiments provided below can refer to the limitations of the artificial intelligence and vehicle dynamics based brake torque distribution method described above, which will not be repeated here.

[0158] In some embodiments, as shown in Figure 8 An artificial intelligence and vehicle dynamics based brake torque distribution device is provided, comprising:

[0159] The acquisition module 11 is configured to acquire the current state information of the target vehicle, the environment information around the target vehicle, and the driving information of the driver on the target vehicle.

[0160] The distribution module 12 is configured to input the state information, the environment information and the driving information into the target torque distribution model to perform torque distribution, and obtain an initial torque distribution result.

[0161] The correction module 13 is configured to correct the initial torque distribution result according to the state information, the driving information and the vehicle dynamics model corresponding to the target vehicle, and obtain a target torque distribution result.

[0162] In some embodiments, the above-mentioned correction module comprises:

[0163] The analysis unit is configured to input the state information and the driving information into the vehicle dynamics model to perform dynamics analysis, and obtain the maximum braking force and the reference yaw rate of the target vehicle.

[0164] The correction unit is configured to correct the initial torque distribution result according to the maximum braking force and the reference yaw rate, and obtain the target torque distribution result.

[0165] In some embodiments, the above-mentioned analysis unit comprises:

[0166] The first analysis sub-unit is configured to input the state information into the braking force analysis model to perform braking force analysis, and obtain the maximum braking force of the target vehicle.

[0167] The second analysis sub-unit is configured to input the state information and the driving information into the angular velocity analysis model to perform analysis, and obtain the reference yaw rate of the target vehicle.

[0168] In some embodiments, the above-mentioned correction unit comprises:

[0169] The first determination subunit is configured to determine whether the initial torque distribution result meets preset braking force conditions and / or preset angular velocity conditions. The preset braking force conditions include that a tire braking force corresponding to the initial braking torque does not exceed a maximum braking force. The preset angular velocity conditions include that an angular velocity difference between a yaw angular velocity corresponding to the initial braking torque and a reference yaw angular velocity does not exceed a preset threshold.

[0170] The second determination subunit is configured to, if the initial torque distribution result meets the preset braking force conditions and the preset angular velocity conditions, take the initial braking torque as the target braking torque.

[0171] The adjustment subunit is configured to, if the initial torque distribution result does not meet the preset braking force conditions or the preset angular velocity conditions, adjust the initial braking torque to obtain an adjusted braking torque, and return to execute the step of determining whether the initial torque distribution result meets the preset braking force conditions and / or the preset angular velocity conditions according to the adjusted braking torque.

[0172] In some embodiments, the adjustment subunit is specifically configured to, if the initial torque distribution result does not meet the preset braking force conditions, take a smaller value between a braking torque corresponding to the maximum braking force and the initial braking torque as the target braking torque; and if the initial torque distribution result does not meet the angular velocity conditions, determine a torque increment according to the angular velocity difference, and adjust the initial braking torque according to the torque increment to obtain the adjusted braking torque.

[0173] In some embodiments, the braking torque distribution device based on artificial intelligence and vehicle dynamics further includes:

[0174] The construction module is configured to construct a simulation scene library of the test vehicle. The simulation scene library includes basic regulation working conditions, extreme danger working conditions, and new energy working conditions.

[0175] The simulation module is configured to control the test vehicle to perform driving tests under various working conditions in the simulation scene library, collect dynamic data of the test vehicle under each working condition, and form a sample data set.

[0176] The training module is configured to train the initial torque distribution model according to the sample data set to obtain a target torque distribution model.

[0177] In some embodiments, the training module includes:

[0178] The decision unit is configured to input the state sample data into an initial decision network to make a braking decision, and output decision action information.

[0179] The prediction unit is configured to input the decision action information and the state sample data into an initial state prediction network to perform state prediction, and obtain state transition information.

[0180] The determining unit is configured to determine a target loss based on the decision action information, the state transition information, and the sample data set.

[0181] The adjusting unit is configured to adjust parameters in the initial decision network based on the target loss until the target loss meets a preset error threshold requirement, and take the trained initial decision network as the target torque distribution model.

[0182] The above-mentioned various modules in the braking torque distribution device based on artificial intelligence and vehicle dynamics can be realized by software, hardware, and combinations thereof, in whole or in part. The above-mentioned various modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned various modules.

[0183] In some embodiments, a computer device is provided, including a memory and a processor, the memory stores a computer program, and the processor implements the steps of the braking torque distribution method based on artificial intelligence and vehicle dynamics according to any of the above-mentioned embodiments when executing the computer program.

[0184] The computer device provided in the above-mentioned embodiments has similar implementation principles and technical effects to the above-mentioned method embodiments, and thus details are not repeated here.

[0185] In some embodiments, a computer readable storage medium is provided, which stores a computer program, and the computer program implements the steps of the braking torque distribution method based on artificial intelligence and vehicle dynamics according to any of the above-mentioned embodiments when executed by a processor.

[0186] The computer readable storage medium provided in the above-mentioned embodiments has similar implementation principles and technical effects to the above-mentioned method embodiments, and thus details are not repeated here.

[0187] In some embodiments, a computer program product is provided, including a computer program, and the computer program implements the steps of the braking torque distribution method based on artificial intelligence and vehicle dynamics according to any of the above-mentioned embodiments when executed by a processor.

[0188] The computer program product provided in the above-mentioned embodiments has similar implementation principles and technical effects to the above-mentioned method embodiments, and thus details are not repeated here.

[0189] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0190] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0191] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A brake torque distribution method based on artificial intelligence and vehicle dynamics, characterized by, The method comprises: acquiring current state information of a target vehicle, environment information around the target vehicle, and driving information of a driver on the target vehicle; inputting the state information, the environment information, and the driving information into a target torque distribution model to perform torque distribution, to obtain an initial torque distribution result; the target torque distribution model is obtained by training based on a pre-constructed simulation scenario library; the simulation scenario library comprises a basic regulation working condition, an extreme danger working condition, and a new energy working condition; the basic regulation working condition comprises a high adhesion straight-line braking working condition, a low adhesion braking working condition, and a steering and braking combined working condition; the extreme danger working condition comprises an opposite open road surface braking working condition, an abutting road surface mutation braking working condition, and a high-speed obstacle avoidance working condition; the new energy working condition comprises a regenerative braking coordination working condition and an energy recovery priority working condition; inputting the state information and the driving information into a vehicle dynamics model to perform dynamics analysis, to obtain a maximum braking force and a reference yaw rate of the target vehicle; determining whether an initial braking torque in the initial torque distribution result meets preset braking force conditions and preset angular velocity conditions; the preset braking force conditions comprise that a tire braking force corresponding to the initial braking torque does not exceed the maximum braking force; the preset angular velocity conditions comprise that an angular velocity difference between a yaw rate corresponding to the initial braking torque and the reference yaw rate does not exceed a preset threshold value; if the initial torque distribution result meets the preset braking force conditions and the preset angular velocity conditions, taking the initial braking torque as a target braking torque in the target torque distribution result; if the initial torque distribution result does not meet the preset braking force conditions, taking a smaller value between the maximum braking force and the initial braking torque as the target braking torque; if the initial torque distribution result does not meet the angular velocity conditions, determining a torque increment according to the angular velocity difference, adjusting the initial braking torque according to the torque increment to obtain an adjusted braking torque, and returning to perform the step of determining whether the initial braking torque in the initial torque distribution result meets the preset braking force conditions and the preset angular velocity conditions according to the adjusted braking torque.

2. The method of claim 1, wherein, The vehicle dynamics model comprises a braking force analysis model and an angular velocity analysis model; the step of inputting the state information and the driving information into the vehicle dynamics model to perform dynamics analysis, to obtain the maximum braking force and the reference yaw rate of the target vehicle, comprises: inputting the state information into the braking force analysis model to perform braking force analysis, to obtain the maximum braking force of the target vehicle; inputting the state information and the driving information into the angular velocity analysis model to perform analysis, to obtain the reference yaw rate of the target vehicle.

3. The method according to claim 1 or 2, characterized in that, The method further comprises: constructing a simulation scenario library of a test vehicle; controlling the test vehicle to perform driving tests under various working conditions in the simulation scenario library, collecting dynamics data of the test vehicle under each working condition, and forming a sample data set; The initial torque distribution model is trained according to the sample data set to obtain the target torque distribution model.

4. The method of claim 3, wherein, The sample data set includes state sample data, action sample data and state transition sample data, the initial torque distribution model includes an initial decision network and an initial state prediction network, and the initial torque distribution model is trained according to the sample data set to obtain the target torque distribution model, including: The state sample data is input into the initial decision network for braking decision, and decision action information is output; The decision action information and the state sample data are input into the initial state prediction network for state prediction, and state transition information is obtained; The target loss is determined according to the decision action information, the state transition information and the sample data set; The parameters in the initial decision network are adjusted according to the target loss until the target loss meets the preset error threshold requirement, and the trained initial decision network is used as the target torque distribution model.

5. An apparatus for AI and vehicle dynamics based brake torque distribution for implementing the AI and vehicle dynamics based brake torque distribution method as claimed in claim 1, characterized in that, The device includes: An acquisition module configured to acquire current state information of a target vehicle, environmental information around the target vehicle, and driving information of a driver on the target vehicle; A distribution module configured to input the state information, the environmental information and the driving information into a target torque distribution model for torque distribution to obtain an initial torque distribution result; A correction module configured to correct the initial torque distribution result according to the state information, the driving information and a vehicle dynamics model corresponding to the target vehicle to obtain a target torque distribution result. 6.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-5 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method of any one of claims 1 to 4.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 4.

Images