Control method suitable for steady-state closed-loop drift of autonomous vehicle

Abstract

The invention discloses a control method suitable for steady-state closed-loop drifting of an automatic driving vehicle, which belongs to the technical field of closed-loop drifting control, and is technically characterized by comprising the following steps of: 1, establishing a whole vehicle dynamics model suitable for controller design; 2, establishing a tire dynamic model; 3, obtaining a drift balance point according to the whole vehicle dynamic model and the tire dynamic model; 4, performing linearization processing on the vehicle model, and performing Taylor expansion on the vehicle model at a drift equilibrium state; 5, the parameters of the yaw velocity, the vehicle speed and the side slip angle of the vehicle are substituted into the error dynamics formula in a unified mode; sixthly, the ideal yaw velocity derivative # imgabs0 # and the ideal course rate # imgabs1 # are substituted into the nonlinear vehicle inverse model; and step 7, outputting a steering wheel angle instruction and a wheel speed instruction through LQR control. The method has the advantages that path tracking and drifting state control are separated through a hierarchical control structure, and the control complexity is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of closed-loop drift control, and specifically relates to a control method suitable for steady-state closed-loop drift of autonomous vehicles. Background Art

[0002] With the continuous development of vehicle intelligence, more and more vehicle controls are gradually pursuing higher maneuverability and stability. The existing related control technologies control the vehicle to be stable within the linear region, making the vehicle in a safer state. In the face of some more extreme working conditions, professional drivers can make more use of vehicle control to achieve more extreme actions. Such as achieving extreme cornering at high speeds through the action of drifting.

[0003] Drifting is a skill that can only be achieved by professional racing drivers. When the vehicle is in an extreme working condition, the coordinated control of the rear-wheel torque and the steering wheel angle brings great driving pleasure. At this time, the vehicle is in an extreme state, and the sideslip angle of the center of mass and the yaw angular velocity are at the "saddle point" position in the phase plane diagram, making the vehicle in a state of losing stability but not losing control.

[0004] For the current drifting working condition, only professional drivers can perform the extreme working condition of drifting. For autonomous vehicles, there is currently no corresponding mass-produced algorithm for this working condition. And it is difficult to achieve the method provided by the current patent for drifting along the corresponding driving trajectory. Summary of the Invention

[0005] Aiming at the deficiencies of the existing technology, the purpose of the embodiments of the present invention is to provide a control method suitable for steady-state closed-loop drift of autonomous vehicles to solve the problems in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions: A control method suitable for steady-state closed-loop drift of autonomous vehicles, comprising the following steps: Step 1: Establish a vehicle dynamics model suitable for controller design, and adopt a three-degree-of-freedom single-track vehicle dynamics model; Step 2: Establish a tire dynamics model. The tire dynamics model adopts the magic formula of tire combined slip, and the corresponding tire parameters are fitted according to the MatLab toolbox to establish the lateral force of the rear wheel , the longitudinal force of the rear wheel , and the lateral force of the front wheel model; Step 3: Based on the above vehicle dynamics model and tire dynamics model, obtain the drift equilibrium point. Combine with the above tire model with slip ratio, and use the Levenberg - Marquardt (LM) algorithm to solve the Fsolve equation by inputting the ideal drift vehicle speed and the steady - state front wheel steering angle; Step 4: Linearize the vehicle model. Perform Taylor expansion on the vehicle model at the drift equilibrium state, neglect the higher - order infinitesimals, obtain the deviation dynamics equation of the system, and separately obtain the Jacobian matrices of the state variables and control variables to complete the linearization of the drifting vehicle model. Use the first - order Euler method to discretize the model and apply the model to the design of the subsequent LQR algorithm; Step 5: According to the specified path information and combined with the actual position of the vehicle, substitute the vehicle's own parameters, yaw rate, vehicle speed, and sideslip angle of the center of mass into the error dynamics formula to obtain the ideal yaw rate derivative , ideal yaw rate , ideal heading rate ; Step 6: Substitute the ideal yaw rate derivative , ideal heading rate into the non - linear vehicle inverse model to obtain the ideal longitudinal force of the rear wheels and the ideal lateral force of the rear wheels , and then substitute them into the rear - wheel thrust angle formula. Obtain the ideal rear - wheel rotational speed by thrust - angle control ; Step 7: Substitute the parameters obtained in Steps 3, 4, 5, and 6 into the linearized state equation to obtain the A and B matrices, and output the steering wheel angle command and wheel speed command through LQR control.

[0007] As a further solution of the present invention, the vehicle dynamics model is shown as the following formula: ; ; ; where is the vehicle's own weight, is the vehicle's longitudinal vehicle speed, is the sideslip angle of the vehicle's center of mass, is the vehicle's yaw rate, is the vehicle's front wheel steering angle, is the vehicle's total moment of inertia, and are the longitudinal force and lateral force of the vehicle's rear wheels, is the distance from the center of mass to the front wheels, is the distance from the center of mass to the rear wheels.

[0008] As a further solution of the present invention, step three specifically includes: Set the state derivative on the left side of the vehicle dynamics model to 0. At this time, the vehicle dynamics model becomes the following system of equations: ; ; ; At the same time, combined with the above tire model with slip ratio, the known longitudinal vehicle speed in the drift equilibrium state is given by the path information, the front wheel steering angle in the drift equilibrium state, an initial solution is set, and the Fsolve equation is solved using the Levenberg-Marquardt (LM) algorithm, and then the sideslip angle of the center of mass in the drift equilibrium state, the yaw rate in the drift equilibrium state, and the rear wheel rotational speed in the drift equilibrium state can be obtained. Then, combined with the above formula, the lateral force of the front wheel in the drift equilibrium state, the lateral force of the rear wheel in the drift equilibrium state, and the longitudinal force of the rear wheel in the drift equilibrium state can be obtained.

[0009] As a further solution of the present invention, step six specifically includes: Substitute the ideal yaw rate derivative and the ideal heading rate into the nonlinear vehicle inverse model; ; ; ; where is the lateral force of the front wheel in the drift equilibrium state, is the lateral force of the rear wheel in the drift equilibrium state, is the front wheel steering angle in the drift equilibrium state, is the longitudinal vehicle speed in the drift equilibrium state, is the yaw rate of the vehicle in the drift equilibrium state.

[0010] As a further solution of the present invention, in step five, there is: ; Then it can be obtained: ; ; The ideal longitudinal force and the ideal rear-wheel lateral force , and then substitute it into the rear-wheel thrust angle formula: ; Obtain the ideal rear-wheel rotational speed by thrust angle control , where is the wheel radius.

[0011] As a further solution of the present invention, the specific steps of step seven include: During the vehicle drifting process, when calculating the rear-wheel torque of the vehicle, the weight roll offset of the vehicle body should be considered. At this time, the ideal rear-wheel rotational speeds of the left and right wheels are different, specifically: ; ; is the vehicle body width, and the mass with roll offset at this time is: ; ; ; Then considering , the torques of the left and right rear wheels can be calculated as: ; ; According to the front-wheel steering angle command and rear-wheel torque command obtained in step seven, control the vehicle to complete closed-loop trajectory following drift.

[0012] As a further solution of the present invention, in step seven, the steering wheel angle command directly controls the steering wheel angle of the vehicle, and the wheel speed command is fused with the current wheel speed of the vehicle for wheel speed control, and finally the rear-wheel torque command is output to control the rear-wheel torque of the vehicle.

[0013] In summary, the embodiments of the present invention have the following beneficial effects compared with the prior art: The upper layer uses error dynamics to control the vehicle trajectory, obtains the drift balance point according to the path information, and provides corresponding parameters to the lower layer; the lower layer uses the LQR algorithm as the vehicle controller, uses the parameters provided by the upper layer path tracking controller and the obtained drift balance point as the controller reference value, and calculates the front-wheel steering angle and rear-wheel driving torque as the vehicle control quantities according to the current vehicle state. The hierarchical control structure separates path tracking and drift state control, reduces the complexity of control, forms a closed-loop drift control, and provides a technical solution for closed-loop drift.

[0014] To more clearly illustrate the structural features and effects of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. Description of the Drawings

[0015] Figure 1 Schematic diagram of the vehicle model for controller design in the present invention Figure 2 Flowchart of a control method applicable to steady-state closed-loop drift of autonomous vehicles in the present invention Figure 3 Schematic diagram of the architecture of the entire algorithm in the present invention Detailed Implementation Manner

[0016] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0017] The following describes the specific implementation of the present invention in detail in conjunction with specific embodiments.

[0018] In one embodiment, a control method applicable to steady-state closed-loop drift of autonomous vehicles, see Figures 1 to 3 , includes the following steps: Step 1: Establish a vehicle dynamics model applicable to controller design, and adopt a three-degree-of-freedom single-track vehicle dynamics model; Step 2: Establish a tire dynamics model. The tire dynamics model adopts the magic formula of combined tire slip, and the corresponding tire parameters are fitted according to the MatLab toolbox to establish the lateral force of the rear wheel , the longitudinal force of the rear wheel , the lateral force of the front wheel model; Step 3: According to the above vehicle dynamics model and tire dynamics model, obtain the drift equilibrium point. Combine with the above tire model with slip ratio, and solve the Fsolve equation using the Levenberg-Marquardt (LM) algorithm by inputting the ideal drift vehicle speed and the steady-state front wheel steering angle; Step 4: Linearize the vehicle model. Perform Taylor expansion on the vehicle model at the drift equilibrium state, ignore the high-order infinitesimals, obtain the deviation dynamics equation of the system, respectively obtain the Jacobian matrices of the state variables and control variables, complete the linearization of the drift vehicle model, and discretize the model using the first-order Euler method for use in the design of the subsequent LQR algorithm; Step 5: According to the specified path information, combine with the actual position of the vehicle, substitute the vehicle's own parameters of yaw rate, vehicle speed, and sideslip angle of the center of mass into the error dynamics formula, and find the ideal derivative of the yaw rate , the ideal yaw rate , Ideal heading rate ; Step 6: Substitute the ideal yaw rate derivative , Ideal heading rate into the non - linear vehicle inverse model to obtain the ideal longitudinal force of the rear wheels and the ideal lateral force of the rear wheels , and then substitute them into the rear - wheel thrust - angle formula. The ideal rear - wheel rotational speed can be obtained by thrust - angle control; Step 7: Substitute the parameters obtained in Steps 3, 4, 5, and 6 into the linearized state equation to obtain the A and B matrices, and output the steering - wheel angle command and wheel - speed command through LQR control.

[0019] Furthermore, referring to Figures 1 to 3 , the vehicle dynamics model is shown as follows: ; ; ; where is the vehicle's own weight, is the vehicle's longitudinal vehicle speed, is the vehicle's center - of - mass sideslip angle, is the vehicle's yaw rate, is the vehicle's front - wheel steering angle, is the vehicle's total moment of inertia, and are the vehicle's rear - wheel longitudinal force and lateral force, is the distance from the center of mass to the front wheels, is the distance from the center of mass to the rear wheels.

[0020] Furthermore, referring to Figures 1 to 3 , Step 3 specifically includes: Set the state derivative on the left - hand side of the vehicle dynamics model to 0. At this time, the vehicle dynamics model becomes the following system of equations: ; ; ; At the same time, combined with the above tire model with slip ratio, the known vehicle longitudinal speed in the drift equilibrium state is given by the path information, the front - wheel steering angle in the drift equilibrium state. Set the initial solution and use the Levenberg - Marquardt (LM) algorithm to solve the Fsolve equation, and the center - of - mass sideslip angle , the yaw rate at the drift equilibrium state , the rear wheel speed at the drift equilibrium state , combined with the above formula, the lateral force of the front wheel at the drift equilibrium state can be obtained and the lateral force of the rear wheel at the drift equilibrium state and the longitudinal force of the rear wheel at the drift equilibrium state .

[0021] Further, referring to Figures 1 to 3 , the specific steps of step six include: Substitute the ideal yaw rate derivative and the ideal heading rate into the non-linear vehicle inverse model; ; ; ; where is the lateral force of the front wheel at the drift equilibrium state, is the lateral force of the rear wheel at the drift equilibrium state, is the front wheel steering angle at the drift equilibrium state, is the longitudinal vehicle speed at the drift equilibrium state, is the yaw rate of the vehicle at the drift equilibrium state.

[0022] Further, referring to Figures 1 to 3 , in step five, there is: ; Then we can get: ; ; Obtain the ideal longitudinal force of the rear wheel and the ideal lateral force of the rear wheel , and then substitute them into the rear wheel thrust angle formula: ; The ideal rear wheel speed is obtained by thrust angle control, where is the wheel radius.

[0023] Further, referring to Figures 1 to 3 , the specific steps of step seven include: During the vehicle drift process, when calculating the rear wheel torque of the vehicle, the weight roll offset of the vehicle body should be considered. At this time, the ideal rear wheel speeds of the left and right wheels are different, specifically: ; ; is the body width of the vehicle, and the mass that undergoes roll offset at this time is: ; ; ; Then considering , the torques of the left and right rear wheels can be calculated as: ; ; According to the front-wheel steering angle command and the rear-wheel torque command obtained in Step Seven, control the vehicle to complete closed-loop trajectory tracking drift.

[0024] Furthermore, referring to Figures 1 to 3 , in Step Seven, the steering wheel angle command directly controls the steering wheel angle of the vehicle, and the wheel speed command is fused with the current wheel speed of the vehicle for wheel speed control, and finally the rear-wheel torque command is output to control the rear-wheel torque of the vehicle.

[0025] In this embodiment, in Step One, a vehicle dynamics model suitable for controller design is established. Due to the controller computing power limitation and the requirement of the vehicle model accuracy, the three-degree-of-freedom single-track vehicle dynamics model is adopted in the present invention.

[0026] ; ; ; Wherein is the self-weight of the vehicle, is the longitudinal vehicle speed, is the sideslip angle of the vehicle's center of mass, is the yaw rate of the vehicle, is the front-wheel steering angle of the vehicle, is the vehicle's moment of inertia, and are the longitudinal and lateral forces of the vehicle's rear wheels, is the distance from the center of mass to the front wheels, is the distance from the center of mass to the rear wheels.

[0027] In Step Two, a tire dynamics model is established. The tire dynamics model adopts the magic formula considering tire combined slip, and the corresponding tire parameters are fitted according to the MatLab toolbox.

[0028] The expression of the ground force on the tire is: ; ; Wherein represents the vertical force received by the rear wheel; and respectively represent the lateral and longitudinal adhesion coefficients of the rear wheel, and are the lateral force and longitudinal force of the rear wheel respectively. At this time, the expression of the vertical acting force of the rear wheel is: ; Assume that the tire has longitudinal and lateral symmetry characteristics. In this case, the longitudinal and lateral adhesion coefficient components are: ; ; ; ; ; where represents the rear wheel rotational speed; represents the rear wheel radius, is the theoretical slip of the tire, is the combined slip of the tire, is the adhesion coefficient calculated using the simplified magic formula, and the expression is: ; where is the tire stiffness factor, is the tire shape factor, is the tire peak factor.

[0029] The rear wheel thrust angle formula is: ; ; Here it is assumed that . Substituting into the formula, we get: ; ; For the front wheel tire force, since the vehicle is rear-wheel drive, there is only a lateral force on the front wheel. At this time, the front wheel side slip angle is: ; The front wheel lateral force applies the standard magic formula as: ; Step 3 According to the above vehicle dynamics model and tire dynamics model, first find the drift balance point. The specific solution method is as follows Let the state derivative on the left side of the vehicle dynamics model be 0. At this time, the vehicle dynamics model becomes an algebraic equation system as follows: ; ; ; Meanwhile, combined with the above tire model with slip ratio, by inputting the ideal drift vehicle speed and the steady-state front wheel angle, and using the Levenberg-Marquardt (LM) algorithm to solve the Fsolve equation, the following can be obtained under the drift equilibrium state , , and then combined with the tire dynamics formula, the following can be obtained under the drift equilibrium state , and .

[0030] Step Four: Since the above method establishes a non-linear system, while the LQR control method is for linear systems, it is necessary to linearize the vehicle model. Perform a Taylor expansion on the vehicle model at the drift equilibrium state, neglect the higher-order infinitesimals, obtain the deviation dynamics equation of the system, and respectively obtain the Jacobian matrices of the state variables and control variables to complete the linearization of the drifting vehicle model. Use the first-order Euler method to discretize the model and apply the model to the design of the LQR algorithm described later.

[0031] The vehicle deviation dynamics equation is as follows: ; where is the deviation of the vehicle state quantity; is the deviation of the control quantity, is the Jacobian matrix of the vehicle dynamics model with respect to the state variables and control variables: ; ; Step Five: First, according to the specified path information and combined with the actual position of the vehicle, define as the distance between the vehicle's center of gravity and the nearest point on the reference path, that is, the lateral error; is the distance along the reference path to this point. Determine the position of the vehicle relative to the reference path through these two quantities. Define as the heading angle of the reference path, as the angle deviation between the vehicle heading and the reference heading.

[0032] The dynamics equation of ; ; Some simplifications are made here, that is , and .

[0033] Therefore, the kinetic equation of ; A stable second-order dynamics is applied to the lateral error and sideslip angle . Combining the above kinetic equation, the desired heading rate is obtained: ; ; ; For the sideslip error , a first-order dynamics is applied to stabilize the sideslip: ; ; The combined yaw rate can be obtained: ; An inner-loop closed loop is formed around the combined yaw rate . By applying a first-order dynamics to the tracking error of the combined yaw rate, the desired first-order yaw acceleration is obtained: ; ; The preset path information is compared with the current path position parameters to obtain the deviation of the current path position from the ideal position and the heading angle deviation : Then, the vehicle's own parameters, yaw rate, vehicle speed, and sideslip angle at the center of mass are uniformly substituted into the error dynamics formula to obtain the ideal yaw rate derivative , the ideal yaw rate , and the ideal heading rate .

[0034] Step six, substitute the ideal yaw rate derivative and the ideal heading rate into the non-linear vehicle inverse model.

[0035] ; ; ; Wherein is the lateral force of the front wheels in the drift equilibrium state, is the lateral force of the rear wheels in the drift equilibrium state, is the steering angle of the front wheels in the drift equilibrium state, is the longitudinal vehicle speed in the drift equilibrium state, is the yaw rate of the vehicle in the drift equilibrium state.

[0036] In step five, there is: ; Then we can get: ; ; Get the ideal longitudinal force of the rear wheels and the ideal lateral force of the rear wheels , and then substitute them into the rear wheel thrust angle formula: ; The ideal rear wheel speed is obtained by thrust angle control, where is the wheel radius.

[0037] In step seven, substitute the parameters obtained in steps three, four, five, and six into the linearized state equation to obtain the A and B matrices, and output the steering wheel angle command and wheel speed command through LQR control.

[0038] Then design the LQR drift controller, where the quadratic optimization function of the LQR drift controller is: ; where and are the weight matrices of the state variables and control variables respectively, and ; ; By solving the equation to obtain , the feedback matrix can be obtained, and the desired control quantity is: ; The steering wheel angle command directly controls the steering wheel angle of the vehicle, and the wheel speed command is fused with the current wheel speed of the vehicle for wheel speed control, and finally the rear wheel torque command is output to control the rear wheel torque of the vehicle.

[0039] During the vehicle drift process, when calculating the rear wheel torque of the vehicle, the weight roll offset of the vehicle body should be considered. At this time, the ideal rear wheel speeds of the left and right wheels are different, specifically: ; ; is the body width of the vehicle, and the mass that undergoes roll offset at this time is: ; ; ; The torques of the left and right rear wheels can be calculated as; ; ; According to the front-wheel steering angle command and the rear-wheel torque command obtained in Step Seven, control the vehicle to complete closed-loop trajectory tracking drift.

[0040] The working principle of the present invention is: It adopts a hierarchical control system structure, integrates a path curvature dynamic update vehicle drift state algorithm, constructs a time-varying transverse and longitudinal coupling dynamics model for complex road drift motion control. The upper layer uses error dynamics to control the vehicle's trajectory, obtains the drift equilibrium point based on path information, and provides corresponding parameters to the lower layer; the lower layer uses the LQR algorithm as the vehicle controller, uses the parameters provided by the upper layer path tracking controller and the obtained drift equilibrium point as the controller reference value, and calculates the front-wheel steering angle and the rear-wheel driving torque as the vehicle control quantities according to the current vehicle state. The hierarchical control structure separates path tracking and drift state control, reduces the complexity of control, forms a closed-loop drift control, and provides a technical solution for closed-loop drift.

[0041] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A control method for steady-state closed-loop drift of an autonomous driving vehicle, characterized in that: The following steps are involved: Step 1: Establish a vehicle dynamics model suitable for controller design, using a three-degree-of-freedom single-track vehicle dynamics model; Step 2: Establish a tire dynamics model. The tire dynamics model uses the magic formula of tire joint slip and fits the corresponding tire parameters according to the MatLab toolbox to establish the rear wheel lateral force. , rear wheel longitudinal force , front wheel lateral force Model; Step 3: According to the above vehicle dynamics model and tire dynamics model, the drift equilibrium point is obtained, and the above tire model with slip rate is combined, and the ideal drift speed and steady-state front wheel angle are input, and the Fsolve equation is solved using the Levenberg-Marquardt (LM) algorithm; Step 4: Linearize the vehicle model, perform Taylor expansion on the vehicle model at the drift equilibrium state, ignore high-order infinitesimals, obtain the system's deviation dynamics equation, obtain the Jacobian matrix of the state variables and the control quantity, complete the linearization of the drift vehicle model, use the first-order Euler method to discretize the model, and use the model in the design of the LQR algorithm described later; Step 5: Based on the specified path information and the actual position of the vehicle, substitute the vehicle's own parameters, yaw rate, vehicle speed, and center of mass sideslip angle, into the error dynamics formula to find the ideal yaw rate derivative. , ideal yaw rate , ideal heading rate ; Step 6: Convert the ideal yaw rate derivative , ideal heading rate Substitute the nonlinear vehicle inverse model to obtain the ideal rear wheel longitudinal force and ideal rear wheel lateral force , and then bring it into the rear wheel thrust angle formula, and get the ideal rear wheel speed by thrust angle control ; Step 7: Substitute the parameters obtained in steps 3, 4, 5 and 6 into the linearized state equation to obtain the A and B matrices, and output the steering wheel angle command and wheel speed command through LQR control.

2. The control method for steady-state closed-loop drift of an autonomous driving vehicle according to claim 1, characterized in that: The vehicle dynamics model is shown as follows: ； ； ； in is the vehicle's own weight, is the longitudinal speed of the vehicle, is the vehicle's center of mass sideslip angle, is the yaw rate of the vehicle, is the front wheel turning angle of the vehicle, is the vehicle moment of inertia, and are the longitudinal and lateral forces on the rear wheels of the vehicle, is the distance from the center of mass to the front wheel, is the distance from the center of mass to the rear wheel.

3. The control method for steady-state closed-loop drift of an autonomous driving vehicle according to claim 1, characterized in that: The step three specifically includes: Let the state derivative on the left side of the vehicle dynamics model be 0. Then the vehicle dynamics model becomes the following set of equations: ； ； ； At the same time, the above tire model with slip rate is combined with the path information to give the known longitudinal speed of the vehicle in the drift equilibrium state. , front wheel turning angle in drift equilibrium state , set the initial solution, use the Levenberg-Marquardt (LM) algorithm to solve the Fsolve equation, and then find the center of mass sideslip angle in the drift equilibrium state , yaw angular velocity in drift equilibrium state , rear wheel speed in drift equilibrium state , combined with the above formula, we can get the lateral force of the front wheel in the drift equilibrium state: , rear wheel lateral force in drift equilibrium state and rear wheel longitudinal force in drift equilibrium state .

4. The control method for steady-state closed-loop drift of an autonomous driving vehicle according to claim 3, characterized in that: The step six specifically includes: The ideal yaw rate derivative , ideal heading rate Substitute the nonlinear vehicle inverse model; ； ； ； in is the lateral force of the front wheel in the drift equilibrium state, is the lateral force of the rear wheel in the drift equilibrium state, is the front wheel turning angle in drift equilibrium state, is the longitudinal speed of the vehicle in drift equilibrium state, is the vehicle yaw rate in drift equilibrium state.

5. The control method for steady-state closed-loop drift of an autonomous driving vehicle according to claim 4, characterized in that: In step five, there are: ； You can get: ； ； Get the ideal rear wheel longitudinal force and ideal rear wheel lateral force , and then bring it into the rear wheel thrust angle formula: ； The ideal rear wheel speed is obtained by thrust angle control ,in is the wheel radius.

6. The control method for steady-state closed-loop drift of an autonomous driving vehicle according to claim 5, characterized in that: The step seven specifically includes: During vehicle drift, the weight roll offset of the vehicle body should be considered when calculating the rear wheel torque of the vehicle. At this time, the ideal rear wheel speeds of the left and right wheels are different, specifically: ； ； is the width of the vehicle, and the mass that causes roll offset is: ； ； ； Consider , the torque of the left and right rear wheels can be calculated as: ； ； According to the front wheel steering angle command and rear wheel torque command obtained in step seven, the vehicle is controlled to complete closed-loop tracking drift.

7. The control method for steady-state closed-loop drift of an autonomous driving vehicle according to claim 6, characterized in that: In step seven, the steering wheel angle command directly controls the steering wheel angle of the vehicle, and the wheel speed command is integrated with the current wheel speed of the vehicle to perform wheel speed control, and finally a rear wheel torque command is output to control the rear wheel torque of the vehicle.

Images