Braking failure working condition vehicle drift collision avoidance control method and system

Abstract

The application discloses a kind of vehicle drift collision avoidance control method and system under brake failure working condition, the method includes: brake failure grade is obtained according to vehicle actual longitudinal braking force parameter and expected longitudinal braking force parameter;Obtain the longitudinal collision risk index of the current state of vehicle according to vehicle brake failure grade;Longitudinal collision risk index is calculated based on high-precision brake failure vehicle dynamics model, to determine failure vehicle drift control strategy according to the index calculation result;Failure vehicle drift control strategy is executed and the corresponding action instruction is output, to send action instruction to vehicle actuator after drift control trigger is executed.The application can be autonomously decided, controlled after vehicle brake failure, to avoid collision accident due to brake failure.Brake failure vehicle can be controlled to change body posture by drift operation, to complete rapid braking and collision avoidance operation by lateral force, to improve the emergency capability and backup level of intelligent vehicle to brake failure.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automobile active safety technology, and in particular to a vehicle drift collision avoidance control method and system under brake failure conditions. BACKGROUND

[0002] Brake failure is a serious threat to traffic safety and the safety of life and property. According to the World Health Organization (WHO), brake failure of vehicles directly causes 2% of traffic accidents, but directly contributes to 8.03% of traffic accident casualties in some areas. The Ministry of Transport of China pointed out that among the factors causing traffic accidents, brake failure accounts for as high as 40%.

[0003] According to statistics, more than 94% of vehicle brake failures are energy failures or failures of the main transmission pressure mechanism, which will cause the braking capacity of all four wheels to fail simultaneously. In the case of simultaneous failure of all four wheels, the maximum longitudinal braking force of the vehicle will decrease sharply, and the longitudinal collision avoidance capability will also decrease significantly, which is prone to cause serious traffic accidents.

[0004] At present, the research on brake failure mainly focuses on redundancy before failure, diagnosis during failure, and control after failure. The above methods can reduce the probability of brake failure events through redundancy, determine the failure cause and failure degree through sensors and intelligent algorithms, and exert the braking performance of the vehicle and stabilize the vehicle body through brake force distribution.

[0005] However, the energy source redundancy and main transmission pressure component redundancy technology in the above measures are still in the development stage, and the simultaneous failure of all four wheels cannot be completely avoided; the existing control measures have conservative requirements for stability after the simultaneous failure of all four wheels, and have not broken through the limitation of longitudinal braking capability, so they cannot improve the forward collision avoidance capability of the whole vehicle. Specifically, if the brake system fails completely, collision is unavoidable without professional driver operation; if the brake system fails partially, the longitudinal braking distance of the vehicle will be significantly lengthened, and forward parking collision avoidance cannot be completed only by brake operation. Accident data caused by brake failure also shows that only longitudinal deceleration control is not the best collision avoidance choice for vehicles with brake failure. In summary, there is still a blank in the collision avoidance safety function of the whole vehicle after the brake system fails.

[0006] Drift operation can control the vehicle to drive by oversteering and side slipping, which is commonly used in racing activities with large changes in road conditions. The essence of drift operation is to control the vehicle to simultaneously achieve oversteering and control of the driving trajectory. Brake failure will cause the limitation of the longitudinal braking force of the vehicle to be unable to be broken through. In this emergency condition, drift operation can be used to change the vehicle body posture and utilize lateral tire force to reduce the vehicle speed.

[0007] Therefore, how to improve the longitudinal collision avoidance capability of the vehicle with brake failure is a problem to be solved at present. SUMMARY

[0008] The present application aims to solve at least one of the technical problems in the related art.

[0009] To this end, the present application aims to provide a vehicle drift collision avoidance control method, system, device and storage medium under brake failure conditions, so as to fully develop the braking potential of the brake failure vehicle and improve the emergency capability and backup level of the intelligent vehicle.

[0010] To achieve the above-mentioned purpose, the present application provides a vehicle drift collision avoidance control method under brake failure conditions, comprising:

[0011] obtaining a brake failure level according to the actual longitudinal braking force parameter and the expected longitudinal braking force parameter of the vehicle;

[0012] obtaining a longitudinal collision risk index under the current state of the vehicle according to the brake failure level of the vehicle;

[0013] calculating the longitudinal collision risk index based on a high-precision brake failure vehicle dynamics model to determine a failure vehicle drift control strategy according to the index calculation result;

[0014] executing the failure vehicle drift control strategy and outputting corresponding action instructions, so as to send the action instructions to the vehicle actuators after the drift control is triggered.

[0015] The vehicle drift collision avoidance control method under brake failure conditions according to the embodiment of the present application can further have the following additional technical features:

[0016] Further, in an embodiment of the present application, before obtaining the brake failure level according to the actual longitudinal braking force parameter and the expected longitudinal braking force parameter of the vehicle, it further comprises: measuring the brake pedal angular displacement by using a brake pedal displacement sensor to calculate the expected longitudinal braking force parameter, and measuring the vehicle acceleration by using an acceleration sensor to calculate the actual longitudinal braking force parameter.

[0017] Further, in an embodiment of the present application, obtaining the longitudinal collision risk index under the current state of the vehicle according to the brake failure level of the vehicle, comprising: obtaining a forward collision risk index by a perception system, and calculating a distance of a forward obstacle from the failure vehicle according to the forward collision risk index; calculating the longitudinal collision risk index under the current state of the vehicle according to the brake failure level and the distance of the forward obstacle from the failure vehicle.

[0018] Further, in an embodiment of the present application, the longitudinal collision risk index is calculated based on the high-precision brake failure vehicle dynamics model to determine the failure vehicle drift control strategy according to the index calculation result, comprising: obtaining the lateral traffic environment parameters according to the state space function and the action space function of the high-precision brake failure vehicle dynamics model in the simulation environment; obtaining the corresponding reward function according to the longitudinal collision risk index and the lateral traffic environment parameters; calculating the corresponding reward function through the high-precision brake failure vehicle dynamics model, and outputting the corresponding failure vehicle drift control strategy according to the function calculation result.

[0019] To achieve the above-mentioned purpose, another aspect of the present application provides a vehicle drift collision avoidance control system under brake failure working condition, comprising:

[0020] The failure level determination module is configured to obtain a brake failure level according to the actual longitudinal braking force parameter and the expected longitudinal braking force parameter of the vehicle.

[0021] The risk index determination module is configured to obtain a longitudinal collision risk index under the current state of the vehicle according to the vehicle brake failure level.

[0022] The control strategy determination module is configured to calculate the longitudinal collision risk index based on the high-precision brake failure vehicle dynamics model to determine the failure vehicle drift control strategy according to the index calculation result.

[0023] The drift control brake module is configured to execute the failure vehicle drift control strategy and output the corresponding action instruction, and send the action instruction to the vehicle actuator after the drift control is triggered.

[0024] The third aspect of the present application provides a computer device comprising a processor and a memory.

[0025] The processor runs the program corresponding to the executable program code in the memory by reading the executable program code stored in the memory, so as to realize the vehicle drift collision avoidance control method under brake failure working condition.

[0026] The fourth aspect of the present application provides a non-transitory computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the vehicle drift collision avoidance control method under brake failure working condition.

[0027] The vehicle drift collision avoidance control method, system, device and storage medium under brake failure working condition of the embodiments of the present application design different drift control strategies for different brake failure levels and lateral traffic environments, and have good scene adaptability and safety.

[0028] The beneficial effects of the embodiments of the present application are:

[0029] 1) The application proposes an active safety function for a brake failure vehicle, which can autonomously decide and control after the vehicle brake failure to avoid collision accidents caused by brake failure.

[0030] 2) The application can control the brake failure vehicle to change the vehicle body posture through drift operation, complete rapid braking and collision avoidance operation by using lateral force, and improve the emergency response capability and backup level of the intelligent vehicle to brake failure.

[0031] 3) The application designs different drift control strategies for different brake failure levels and lateral traffic environments, which has good scene adaptability and safety.

[0032] Additional aspects and advantages of the application will be in part apparent and in part pointed out hereinafter in the description. BRIEF DESCRIPTION OF DRAWINGS

[0033] The above and / or additional aspects and advantages of the application will become apparent and be readily understood from the following description, taken in conjunction with the accompanying drawings, in which:

[0034] Figure 1 The flow chart of the vehicle drift collision avoidance control method under brake failure working condition according to the embodiment of the application is shown in the figure.

[0035] Figure 2 The architecture diagram of the vehicle drift collision avoidance control under brake failure working condition according to the embodiment of the application is shown in the figure.

[0036] Figure 3 The accuracy verification effect diagram of the vehicle dynamics model according to the embodiment of the application is shown in the figure.

[0037] Figure 4 The tire model fitting effect diagram according to the embodiment of the application is shown in the figure.

[0038] Figure 5 The comparison diagram of the accuracy of the vehicle dynamics model under extreme working condition according to the embodiment of the application is shown in the figure.

[0039] Figure 6 The vehicle G-G diagram after brake failure according to the embodiment of the application is shown in the figure.

[0040] Figure 7 The Critic network schematic diagram according to the embodiment of the application is shown in the figure.

[0041] Figure 8 The Actor network schematic diagram according to the embodiment of the application is shown in the figure.

[0042] Figure 9 The comparison diagram of the output vehicle braking trajectory and the failure vehicle trajectory output only by braking according to the embodiment of the application is shown in the figure.

[0043] Figure 10 Tire force utilization effect diagram under the control strategy according to the embodiment of the application;

[0044] Figure 11 Structural schematic diagram of the vehicle drift collision avoidance control system under the brake failure condition according to the embodiment of the application;

[0045] Figure 12 Computer device according to the embodiment of the application. DETAILED DESCRIPTION

[0046] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0047] In order for those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the protection scope of the present application.

[0048] The vehicle drift collision avoidance control method, system, device and storage medium under the brake failure condition according to the embodiments of the present application will be described below with reference to the drawings.

[0049] Figure 1 is a flowchart of the vehicle drift collision avoidance control method under the brake failure condition according to an embodiment of the present application.

[0050] As shown in Figure 1 , the method includes but is not limited to the following steps:

[0051] S1, obtaining a brake failure grade according to an actual longitudinal braking force parameter and an expected longitudinal braking force parameter of the vehicle.

[0052] S2, obtaining a longitudinal collision risk index in a current state of the vehicle according to the brake failure grade of the vehicle;

[0053] S3, calculating the longitudinal collision risk index based on a high-precision brake failure vehicle dynamics model to determine a failure vehicle drift control strategy according to a calculation result of the index;

[0054] S4, executing the failure vehicle drift control strategy and outputting a corresponding action instruction to send the action instruction to a vehicle actuator after the drift control is triggered.

[0055] The method of the embodiments of the present application will be described in detail below with reference to the drawings.

[0056] The method of the embodiment of the application can be divided into constructing a vehicle brake failure level observation module, establishing a longitudinal space-time risk assessment model of a brake failure vehicle, and outputting a drift control strategy based on reinforcement learning combined with a lateral traffic environment. In addition, the failure drift control strategy also includes other automatic drift control strategies. The specific implementation is as shown in Figure 2 .

[0057] Specifically, the vehicle brake failure level observation module includes: a brake pedal displacement sensor measuring a brake pedal angular displacement, and a table lookup calculating an expected longitudinal braking force at this time; an acceleration sensor measuring a vehicle deceleration, and calculating an actual longitudinal braking force at this time. The brake failure level is divided into four levels of energy failure (91%-100%), extreme failure (76%-90%), serious failure (51%-75%), partial failure (26%-50%), and slight failure (10%-25%), and the brake failure level is determined according to the vehicle actual longitudinal braking force and the expected longitudinal braking force.

[0058] As an example, during vehicle driving, the pedal sensor returns an angular displacement α b , and a table lookup obtains an expected braking force F(α b ) at this time; at this time, the vehicle acceleration sensor returns a longitudinal acceleration a x . The vehicle brake failure level can be calculated according to the following formula, and in this embodiment, the failure level is extreme failure (failure 80%).

[0059]

[0060] Further, the longitudinal space-time collision risk assessment module of the brake failure vehicle includes: analyzing a forward collision risk through a perception system of an automatic driving system, and measuring a distance of a forward obstacle from the failure vehicle. In combination with the failure level output by the failure level observation module, the longitudinal collision risk of the vehicle in the current state is calculated. According to the longitudinal collision risk, it is judged whether to execute drift collision avoidance control.

[0061] As an example, the automatic driving perception system feeds back that an obstacle appears at a forward l obs of the vehicle, and the longitudinal collision avoidance demand increases sharply. The space-time collision risk R ego of the failure vehicle in the current state is calculated.

[0062]

[0063] When R ego is greater than 1, the drift control strategy of the application will be triggered. In this embodiment, the drift control strategy is triggered.

[0064] Further, the failure vehicle drift control strategy based on reinforcement learning includes: constructing a state space and an action space of a SAC deep reinforcement learning algorithm; establishing a high-precision brake failure vehicle dynamics model as an environment of the SAC reinforcement learning algorithm; designing a reward function according to a longitudinal space-time collision risk and a lateral traffic environment, including an immediate reward R i and a terminal reward R T ; training network parameters of the reinforcement learning algorithm in a simulation environment, and importing the network parameters into a vehicle-mounted computer in a real environment; after the drift control is triggered, the vehicle-mounted computer sends an action instruction to a vehicle actuator to realize drift braking. Specifically:

[0065] The state space S includes information required for intelligent vehicle collision avoidance under brake failure conditions, and the space S includes information required for intelligent vehicle collision avoidance under brake failure conditions, including the state of the ego vehicle and the surrounding longitudinal and lateral environment;

[0066] The action space includes all controllable quantities of the intelligent vehicle under brake failure conditions, including but not limited to front and rear wheel steering angles, and wheel braking / driving torques;

[0067] The high-precision brake failure vehicle dynamics model can accurately calculate the state transition of the intelligent vehicle after brake failure, including but not limited to lateral, longitudinal and vertical displacements, yaw and roll.

[0068] The reward function includes a terminal reward R T and an immediate reward R i . The terminal reward is a one-time reward given at the end of each training round according to the terminal state, including: an undamaged reward term R T1 , a rollover penalty term R T2 , a collision or exceeding the lateral road boundary penalty term R T3 . The immediate reward is an immediate reward given after each state transition, including: a collision avoidance term R i1 , for rewarding the vehicle deceleration value in the world coordinate system; a deceleration term R i2 , for rewarding the absolute deceleration of the vehicle; a lateral displacement term R i3 , for punishing the lateral displacement of the vehicle in the world coordinate system; a tire friction utilization term R i4 , for indicating the proximity of the vehicle motion state to the current extreme limit dynamic performance.

[0069] The lateral displacement penalty term R i2 needs to be determined whether to suppress the lateral displacement of the brake failure vehicle according to the lateral road traffic environment. If the lateral displacement does not need to be suppressed, the penalty term can be downgraded or removed.

[0070] The tire friction utilization term R i4, the current available tire force range needs to be determined according to the brake failure level, and the angle beta between the maximum tire force and the vehicle body is calculated.

[0071] In addition, the failure drift control strategy also includes other automatic drift control strategies.

[0072] As an example, the control strategy of the current failure level vehicle is designed based on reinforcement learning. The state space S includes the information required for the intelligent vehicle to avoid collision under the brake failure working condition, and the state space of the embodiment is as follows:

[0073] S=[s ego ,s enr ]

[0074] s ego =[v x ,v y ,r,γ,X e ,Y e ,ψ,φ,s]

[0075]

[0076] In the formula, s ego ,s enr are the required self-vehicle state information and surrounding environment information in the embodiment; v x ,v y ,r,γ are the longitudinal speed, lateral speed, yaw rate and roll speed of the vehicle in the vehicle coordinate system; X e ,Y e ,ψ,φ are the longitudinal position, lateral position, yaw angle and roll angle of the vehicle in the world coordinate system; s is the state of the failure vehicle, including: collision avoidance failure, collision avoidance success, collision avoidance process, and roll; is the deceleration of the vehicle in the world coordinate system in the direction of the connecting line with the obstacle.

[0077] Further, the vehicle studied in the embodiment of the application is a four-wheel steering four-wheel drive automobile, and the action space A is constructed as shown below:

[0078]

[0079] In the formula, δ f ,δ r are the steering angles of the front wheels and the rear wheels; is the drive / brake torque of the four wheels under the failure condition.

[0080] Further, the embodiment of the application constructs a high-precision eight-degree-of-freedom brake failure vehicle dynamics model as a simulation environment, and the comparison between the model and the Carsim vehicle model is shown in Figure 3 .

[0081] Wherein, the vehicle body model is a double-track four-degree-of-freedom vehicle body model:

[0082]

[0083] Wherein, F x1 ,F x2 ,F x3 ,F x4 is the longitudinal force provided by the four-wheel tires; F y1 ,F y2 ,F y3 ,F y4 is the lateral force provided by the four-wheel tires; L a ,L b are the distance from the front axle to the center of mass and the distance from the rear axle to the center of mass, respectively; W f ,W r are the front wheel track and the rear wheel track, respectively; and h is the vehicle center of mass height.

[0084]

[0085] M x = mgh phi + (-D f gamma - K f phi) + (-D r gamma - K r phi)

[0086] Wherein, rho is the air density, C d is the air resistance coefficient, and A is the vehicle cross-sectional area; D f ,D r are the front roll damping and the rear roll damping, respectively; K f ,K r are the front roll stiffness and the rear roll stiffness, respectively.

[0087] Further, the tire model of the embodiment of the present application adopts a magic formula fitted by experimental data, and the accuracy of the tire model is as shown in Figure 4 .

[0088]

[0089]

[0090]

[0091]

[0092] Wherein, S x is the longitudinal slip ratio, alpha is the tire side slip angle, and the remaining variables are fitting parameters.

[0093] Further, the drift control of the embodiment of the present application relates to limit driving, the tire side angle will be obviously higher than the normal driving scene, so the side angle needs to be corrected. The corrected vehicle model can provide a limit working condition simulation environment with high tire side angle, and the vehicle model is compared with the Carsim output result as shown in the following table. Figure 5 .

[0094] α = π - αif

[0095] α = - π + αif

[0096] α = αif

[0097] Further, the design of the reward function of the embodiment of the present application: the setting of the reward function includes the terminal reward R T and the immediate reward R i .

[0098] R = R i + R T

[0099] The immediate reward includes: the collision avoidance term R i1 , used for rewarding the vehicle deceleration value in the world coordinate system; the deceleration term R i2 , used for rewarding the absolute deceleration of the vehicle; the lateral displacement term R i3 , used for punishing the lateral displacement of the vehicle in the world coordinate system; the tire friction utilization term R i4 , used for representing the proximity of the vehicle motion state to the current limit driving performance. The specific description is as follows:

[0100] The collision avoidance term R i1 , rewards the deceleration of the vehicle in the direction of the line connecting the vehicle and the obstacle in the world coordinate system. The greater the deceleration represents the degree of immediate collision avoidance effort of the failed vehicle.

[0101]

[0102] The deceleration term R i2 , rewards the absolute deceleration of the vehicle. The deceleration represents the reduction of the absolute speed, and encourages the absolute speed to decelerate to reduce the potential collision loss.

[0103]

[0104] The lateral displacement term R i3 , punishes the lateral displacement of the vehicle in the world coordinate system;

[0105]

[0106] In the formula, yla-obs is the relative lateral position of the obstacle and the ego vehicle in the lateral traffic environment. If there is no obstacle in the lateral direction, the lateral displacement term can be degraded or removed.

[0107] tire friction utilization term R i4 , which rewards the proximity of the vehicle motion state to the current limit dynamic performance. This term needs to determine the range of current available tire force according to the brake failure level, and calculate the angle β between the maximum tire force and the vehicle body, as shown in Figure 6 .

[0108]

[0109] The final reward includes: the completion of the collision avoidance reward term R T1 , the occurrence of the roll penalty term R T2 , the occurrence of the collision or exceeding the lateral road boundary penalty term R T3 .

[0110]

[0111] In the formula, k5 is a positive number, and a high order reward is given if no collision occurs; k6 and k7 are negative numbers, and a high order penalty is given if roll, collision or road boundary is exceeded.

[0112] Further, the embodiment of the present application constructs the Critic and Actor deep neural networks of the reinforcement learning algorithm, as shown in Figure 7 , Figure 8 The reward R is determined according to the state S and the action A, and the Critic network parameters are trained according to the minimization of the time difference loss function; the Actor network parameters are trained by maximizing the value function.

[0113] Further, the embodiment of the present application imports the aforementioned trained network parameters into the vehicle-mounted computer in the real environment; after the drift control trigger, the vehicle-mounted computer sends the action instruction to the vehicle actuator to realize drift braking. In this embodiment, the braking trajectory of the vehicle is as shown in Figure 9 , and the tire force is as shown in Figure 10 .

[0114] The vehicle drift collision avoidance control method under brake failure working condition according to the embodiment of the present application can autonomously decide and control after the vehicle brake failure, and avoid collision accidents caused by brake failure. The brake failure vehicle can be controlled to change the vehicle body posture through drift operation, and utilize the lateral force to complete rapid braking and collision avoidance operation, thereby improving the emergency response capability and backup level of the intelligent vehicle to brake failure. Different drift control strategies are designed for different brake failure levels and lateral traffic environments, and have good scene adaptability and safety.

[0115] In order to realize the above-mentioned embodiments, asFigure 11 In the embodiment, a vehicle drift collision avoidance control system 10 under brake failure condition is also provided, which comprises a failure level determination module 100, a risk index determination module 200, a control strategy determination module 300 and a drift control module 400.

[0116] The failure level determination module 100 is configured to obtain a brake failure level according to an actual longitudinal braking force parameter and an expected longitudinal braking force parameter of the vehicle.

[0117] The risk index determination module 200 is configured to obtain a longitudinal collision risk index under a current state of the vehicle according to the brake failure level of the vehicle.

[0118] The control strategy determination module 300 is configured to calculate the longitudinal collision risk index based on a high-precision brake failure vehicle dynamics model, and determine a failure vehicle drift control strategy according to a calculation result of the index.

[0119] The drift control module 400 is configured to execute the failure vehicle drift control strategy and output a corresponding action instruction, and send the action instruction to a vehicle actuator after the drift control is triggered.

[0120] Further, before the failure level determination module 100, a parameter calculation module is further included.

[0121] The parameter calculation module is configured to measure a brake pedal angular displacement by using a brake pedal displacement sensor to calculate the expected longitudinal braking force parameter, and measure a vehicle acceleration by using an acceleration sensor to calculate the actual longitudinal braking force parameter.

[0122] Further, the risk index determination module 200 is further configured to:

[0123] obtain a forward collision risk index by using a perception system, and calculate a distance from a failure vehicle to a forward obstacle according to the forward collision risk index.

[0124] obtain the longitudinal collision risk index under the current state of the vehicle according to the brake failure level and the distance from the failure vehicle to the forward obstacle.

[0125] Further, the control strategy determination module 300 is further configured to:

[0126] obtain a lateral traffic environment parameter according to a state space function and an action space function of the high-precision brake failure vehicle dynamics model in a simulation environment.

[0127] obtain a corresponding reward function according to the longitudinal collision risk index and the lateral traffic environment parameter.

[0128] The corresponding reward function is calculated through a high-precision brake failure vehicle dynamics model, and a corresponding failure vehicle drift control strategy is output according to the calculation result of the function.

[0129] The vehicle drift collision avoidance control system under the brake failure working condition according to the embodiment of the application can autonomously make decisions and control after the vehicle brake failure, and avoid collision accidents caused by brake failure. The brake failure vehicle can be controlled to change the vehicle body posture through drift operation, and rapid braking and collision avoidance operation can be completed by using lateral force, thereby improving the emergency response capability and backup level of the intelligent vehicle to brake failure. Different drift control strategies are designed for different brake failure levels and lateral traffic environments, and have good scene adaptability and safety.

[0130] In order to realize the method of the above-mentioned embodiments, the application further provides a computer device, as shown in the accompanying drawings, the computer device 600 comprises a memory 601, a processor 602; wherein the processor 602 runs the program corresponding to the executable program code by reading the executable program code stored in the memory 601, so as to realize each step of the vehicle drift collision avoidance control method under the brake failure working condition described above. Figure 12

[0131] In order to realize the method of the above-mentioned embodiments, the application further provides a non-temporary computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the vehicle drift collision avoidance control method under the brake failure working condition.

[0132] In addition, the terms "first", "second", "third" and the like are used only to describe various features, and are not to be construed as indicating or implying relative importance or a specific number of the features indicated. Therefore, a feature defined with "first", "second", "third" or the like can explicitly or implicitly include at least one of the features. In the description of the application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise specifically limited.

[0133] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the description of the present application, the illustrative description of the above terms is not necessarily for the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any suitable manner in any one or more embodiments or examples. In addition, the skilled in the art can combine and combine the different embodiments or examples described in the present application and the features of the different embodiments or examples without contradiction.

[0134] ​Although the embodiments of the present application have been shown and described above, it is understood that the above-described embodiments are exemplary and are not to be construed as limiting the present application, and that variations, modifications, substitutions and changes can be made by those skilled in the art without departing from the scope of the present application.

Claims

1. A method for vehicle drift collision avoidance control in brake failure condition, characterized in that, The method comprises the following steps: obtaining a brake failure level according to the actual longitudinal braking force parameter and the expected longitudinal braking force parameter of the vehicle; calculating a longitudinal collision risk index in the current state of the vehicle according to the brake failure level of the vehicle; determining a failure vehicle drift control strategy according to the longitudinal collision risk index; executing the failure vehicle drift control strategy and outputting corresponding action instructions, so that the action instructions are sent to the vehicle actuators after the drift control is triggered; the failure vehicle drift control strategy is a failure vehicle drift control strategy based on reinforcement learning, comprising: constructing a state space and an action space of a SAC deep reinforcement learning algorithm; Establish high-precision dynamics model of brake failure vehicle as the environment of SAC reinforcement learning algorithm; design reward function according to longitudinal collision risk index and lateral traffic environment, including immediate reward and terminal reward ; train network parameters of reinforcement learning algorithm in simulation environment, and import network parameters into vehicle-mounted computer in real environment.

2. The method of claim 1, wherein, before the step of obtaining the brake failure level according to the actual longitudinal braking force parameter and the expected longitudinal braking force parameter of the vehicle, the method further comprises the following steps: measuring the angular displacement of the brake pedal by using a brake pedal displacement sensor to calculate the expected longitudinal braking force parameter, and measuring the acceleration of the vehicle by using an acceleration sensor to calculate the actual longitudinal braking force parameter.

3. The method of claim 1, wherein, calculating a longitudinal collision risk index in the current state of the vehicle according to the brake failure level of the vehicle, comprising: calculating a longitudinal collision risk index in the current state of the vehicle according to the brake failure level and the distance of the forward obstacle from the failure vehicle.

4. The method of claim 1, wherein, determining a failure vehicle drift control strategy according to the longitudinal collision risk index, comprising: obtaining lateral traffic environment parameters according to the state space function and the action space function of the high-precision brake failure vehicle dynamics model in the simulation environment; obtaining a corresponding reward function according to the longitudinal collision risk index and the lateral traffic environment parameters; calculating the corresponding reward function by using the high-precision brake failure vehicle dynamics model, and outputting a corresponding failure vehicle drift control strategy according to the function calculation result.

5. A vehicle drift avoidance control system in a brake failure condition, characterized by, comprising: a failure level determination module for obtaining a brake failure level according to the actual longitudinal braking force parameter and the expected longitudinal braking force parameter of the vehicle; a risk index determination module for calculating a longitudinal collision risk index in the current state of the vehicle according to the brake failure level of the vehicle; a control strategy determination module for determining a failure vehicle drift control strategy according to the longitudinal collision risk index; a drift control module for executing the failure vehicle drift control strategy and outputting corresponding action instructions, so that the action instructions are sent to the vehicle actuators after the drift control is triggered; the failure vehicle drift control strategy is a failure vehicle drift control strategy based on reinforcement learning, comprising: constructing a state space and an action space of a SAC deep reinforcement learning algorithm; Establish high-precision dynamics model of brake failure vehicle as the environment of SAC reinforcement learning algorithm; design reward function according to longitudinal collision risk index and lateral traffic environment, including immediate reward and terminal reward ; train network parameters of reinforcement learning algorithm in simulation environment, and import network parameters into vehicle-mounted computer in real environment.

6. The system of claim 5, wherein, further comprising: a parameter calculation module, for measuring the angular displacement of the brake pedal by using a brake pedal displacement sensor to calculate the expected longitudinal braking force parameter, and measuring the acceleration of the vehicle by using an acceleration sensor to calculate the actual longitudinal braking force parameter.

7. The system of claim 5, wherein, the risk index determination module is further used for: calculating a longitudinal collision risk index in the current state of the vehicle according to the brake failure level and the distance of the forward obstacle from the failure vehicle.

8. The system of claim 5, wherein, the control strategy determination module is further used for: obtaining lateral traffic environment parameters according to the state space function and the action space function of the high-precision brake failure vehicle dynamics model in the simulation environment; According to the longitudinal collision risk index and the lateral traffic environment parameter, a corresponding reward function is obtained; The corresponding reward function is calculated through the high-precision brake failure vehicle dynamics model, and a corresponding failure vehicle drift control strategy is output according to the function calculation result.

9. A computer device, comprising: comprise a processor and a memory; The processor executes the executable program code stored in the memory to run the program corresponding to the executable program code, so as to realize the vehicle drift collision avoidance control method under brake failure working condition as claimed in any one of claims 1-4.

10. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to realize the vehicle drift collision avoidance control method under brake failure working condition as claimed in any one of claims 1-4.

Images