An emergency obstacle avoidance path planning and control method based on model predictive control

Abstract

The present invention discloses an emergency obstacle avoidance path planning and control method based on model predictive control, including a decision-making part, a path planning part and a tracking control part, which are as follows: a decision-making part: when an obstacle is detected during high-speed operation of a vehicle and it is determined that the obstacle cannot be avoided by conventional turning, a decision to perform drift obstacle avoidance is made; a path planning part: after the decision to perform drift obstacle avoidance is made, a forward drift obstacle avoidance path is searched according to the current vehicle speed to generate a drift obstacle avoidance path; wherein the forward drift obstacle avoidance path search is based on a hybrid path search and an optimized path search method; a tracking control part: based on a linearized vehicle dynamics model, a model predictive controller with a balance point during drifting as a state quantity is used to track the generated drift obstacle avoidance path. By using the present invention, effective obstacle avoidance can be performed when a high-speed vehicle faces a sudden obstacle.

Description

Technical Field

[0001] The present invention relates to the field of unmanned vehicle path planning and path tracking, and in particular to an emergency obstacle avoidance path planning and control method based on model predictive control. Background Art

[0002] Steady-state drift is a typical extreme vehicle driving condition. "Steady-state" means that the state of the vehicle system (including vehicle speed, yaw angle, sideslip angle, etc.) remains unchanged during the drift process. Experienced racing drivers usually force the vehicle into a steady-state drift state by coordinating the rear axle driving force and the front wheel turning angle to shorten the cornering time.

[0003] With the development of the automobile industry and the continuous maturity of system theory, automobiles are moving towards intelligence and unmanned driving. The emergency avoidance capability and adaptability to various working conditions of unmanned vehicles have attracted widespread attention and become a new research direction for autonomous driving.

[0004] When a car is drifting, it is in an extreme condition and can quickly change its direction to avoid or mitigate collisions. Therefore, improving the emergency avoidance capability of autonomous vehicles by studying drift control is an important research direction. However, the current vehicle drift control has the following problems:

[0005] There is strong nonlinearity, which makes it very difficult to model the vehicle drift process, and there is complex coupling between the lateral and longitudinal directions, making control difficult: when drifting, the rear wheels are at the adhesion circle limit, and the lateral and longitudinal forces of the rear wheels are constrained by the adhesion circle limit. At this time, there is a complex coupling relationship between the lateral and longitudinal directions of the vehicle. Most of the current technical solutions use a combination of open-loop and closed-loop control, introducing the experience of professional drivers for open-loop control. This solution has weak anti-interference ability and requires reliance on experience for controller design.

[0006] A Chinese patent document with publication number CN113928311A discloses a closed-loop switching control method for steady-state drift of a vehicle, comprising: S1. calculating the steady-state drift domain and steady-state drift equilibrium point of the vehicle, and decomposing the drift process of the vehicle into a transition stage approaching the steady-state drift domain and a steady-state stage entering the steady-state drift domain based on the steady-state drift domain; S2. designing a transition drift controller for the transition stage based on a two-degree-of-freedom vehicle model, and controlling the vehicle to approach the steady-state drift domain based on the transition drift controller, wherein the transition drift controller comprises a longitudinal PID controller and a lateral LQR controller; S3. judging whether the current state quantity deviation of the vehicle enters the steady-state drift domain, if yes, entering step S4, otherwise returning to step S2; S4. designing a steady-state drift LQR controller for the steady-state stage based on a three-degree-of-freedom vehicle model, and controlling the vehicle to track the steady-state drift equilibrium point.

[0007] The above method can avoid the influence of open-loop control on the anti-interference ability of the system, make the steady-state drift control scheme repeatable, and can realize the steady-state drift of the vehicle, which is suitable for vehicle control research under extreme working conditions. However, this method can only solve the repeatability of steady-state drift control, and cannot effectively avoid obstacles when facing complex and sudden situations, such as when a high-speed vehicle (close to 100Km / h) faces a sudden obstacle. Summary of the invention

[0008] The present invention provides an emergency obstacle avoidance path planning and control method based on model predictive control, which can effectively avoid obstacles when a high-speed vehicle (close to 100 km / h) faces obstacles that suddenly appear.

[0009] An emergency obstacle avoidance path planning and control method based on model predictive control includes a decision-making part, a path planning part and a tracking control part, as follows:

[0010] Decision-making part: When the vehicle detects an obstacle during high-speed operation and determines that it cannot avoid the obstacle through conventional turns, it makes a decision to drift to avoid the obstacle;

[0011] Path planning part: After the decision to drift and avoid obstacles is made, a forward drift and avoidance path is searched according to the current vehicle speed to generate a drift and avoidance path; the forward drift and avoidance path search is based on a hybrid path search and an optimized path search method;

[0012] Tracking control part: Based on the linearized vehicle dynamics model, a model predictive controller with the equilibrium point during drifting as the state quantity is used to track the generated drift obstacle avoidance path.

[0013] In the decision-making part, after detecting an obstacle, the vehicle determines whether the minimum turning radius can be used to avoid the obstacle based on the obstacle position and the current vehicle speed. If it is determined that the minimum turning radius cannot effectively avoid the obstacle, a decision is made to use drifting to avoid the obstacle.

[0014] In the path planning part, the hybrid path search and optimized path search method uses the front wheel angle to perform forward path search based on vehicle dynamics; for each front wheel angle search point, a search weight calculation based on the optimization problem is performed, and the search point with the largest weight is selected as the search point for forward path planning based on the weight of each search point. After the forward output of each point is optimized, the system model state and input value at each moment are calculated as a reference for the system state and input of the model predictive controller at each moment in the tracking control part.

[0015] The front wheel turning angle is used as the search point. When the vehicle drifts counterclockwise, the front wheel turning angle takes a value of -10°:-10°:-30°. When the vehicle drifts clockwise, the front wheel turning angle takes a value of 10°:10°:30°.

[0016] The generated drift obstacle avoidance path satisfies both obstacle avoidance requirements and vehicle drift dynamics. The formula for vehicle drift dynamics is as follows:

[0017]

[0018]

[0019]

[0020] in, is the derivative of velocity, is the derivative of the vehicle sideslip angle, is the derivative of the vehicle's yaw rate, V is the vehicle speed, β is the vehicle's sideslip angle, r is the vehicle's yaw rate, δ is the front wheel turning angle, a and b are the distances from the vehicle's center of mass to the front and rear wheels, respectively, and F yf and F yr are the lateral forces on the front and rear wheels of the vehicle, respectively, and F xr is the rear wheel driving force of the vehicle, m is the vehicle mass, I z is the moment of inertia of the vehicle around the Z axis; The equation at the drift equilibrium point is obtained, and the values ​​of various parameters at the drift equilibrium point are obtained after solving V eq ,β eq ,r eq ,δ eq , Among them, β eq and V eq is the value of the sideslip angle and vehicle speed during steady-state drift, r eq represents the expected value of the yaw rate when the vehicle drifts, δ eq and are the values ​​of the front wheel steering angle and rear wheel driving force at the equilibrium point respectively.

[0021] According to the value of the vehicle drift balance point, a heuristic search function is used to forward fit the drift obstacle avoidance path. The heuristic search function takes the form of the following optimization function:

[0022]

[0023] st

[0024] Δξ(k+1)=AΔξ(k)+BΔu(k)

[0025] Δu min ≤Δu(k)≤Δumax

[0026] In the formula, r k represents the predicted yaw rate at time k, r eq represents the expected value of the yaw rate when the vehicle drifts; similarly, β k and V k They represent the predicted value of the vehicle sideslip angle and the predicted value of the vehicle body speed at time k respectively; β eq and V eq is the value of the sideslip angle and vehicle speed during steady-state drift; δ k and Fxr k are the predicted values ​​of the front wheel steering angle and rear wheel driving force at time k, δ eq and are the values ​​of the front wheel steering angle and rear wheel driving force at the equilibrium point respectively; Δx(k) and Δu(k) are the deviations of the system state quantity and the input quantity relative to the equilibrium point value at time k respectively, and the system state quantity is ξ=[r,β,V] T , the system input is u=[δ,F xr ] T .

[0027] After solving the optimization problem, the drift path is calculated. The formula is as follows:

[0028]

[0029]

[0030]

[0031]

[0032] For each front wheel turning angle δ, there is a corresponding forward search path [(X1, Y1), (X2, Y2), …, (X n ,Y n )]; For each front wheel turning angle corresponding to the forward search path, the weight corresponding to the current front wheel turning angle value is calculated, and the weight is expressed by the following formula:

[0033]

[0034] (X obs ,Y obs ) is the obstacle coordinate; according to the weight value H corresponding to each front wheel search angle, the largest value is selected as the drift obstacle avoidance path.

[0035] The model predictive controller adopts the model predictive control framework. The model in the model predictive control framework adopts a linear discrete model of dynamics, which is obtained by linearizing and discretizing the dynamic equations:

[0036] Δx(k+1)=AΔx(k)+BΔu(k)

[0037] In the formula, x(k) is the system state, u(k) is the system input, the linearization method uses the Taylor expansion method, and the discretization method uses the forward Euler method; in order to track the planned path, the state x = [e, Δφ, V, β, r] is selected T , input u=[δ,F xr ] T , e is the lateral error of the vehicle's current position from the reference path, Δφ is the deviation between the current heading angle and the desired heading angle, V is the vehicle's center of mass velocity, β is the center of mass sideslip angle, and r is the vehicle's yaw rate; the input δ is the front wheel steering angle, F xr is the rear wheel driving force; the system matrix A and input matrix B are obtained by

[0038]

[0039]

[0040] In the above matrix, y1,y2,…,y n (n=5) is the following equation:

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] After the system model is proposed, it is necessary to construct the optimization problem of the model predictive controller (MPC) and solve the optimization problem in the form of solving the QP problem. The model predictive controller (MPC) optimization problem needs to optimize the model input at the time (k~k+Np-1). The optimization function is the following expression:

[0047]

[0048] st

[0049] Δx(k+1)=AΔx(k)+BΔu(k)

[0050] Δu min ≤Δu(k)≤Δu max

[0051] x(k) is the input of the model at time k x = [e, Δφ, V, β, r] T ,x eq is the value of the equilibrium point of the drift process model state quantity, u(k) is the model input quantity at time k u=[δ,F xr ] T The value of Δu min and Δu max are the minimum and maximum allowable input deviations at time k and time k-1, and A and B are the system matrix and input matrix obtained after linearization + discretization, respectively.

[0052] The method of the present invention can adopt a drifting obstacle avoidance strategy to avoid obstacles when the vehicle determines that conventional turns cannot effectively avoid obstacles when obstacles appear. When the vehicle detects the appearance of obstacles, it can immediately plan the path ahead, generate the forward tracking path by using a hybrid search and optimization method, establish a model predictive controller based on drift dynamics, and use the model predictive control method to track the path generated in the previous step to complete the drifting obstacle avoidance work. The purpose of avoiding obstacles that suddenly appear can be achieved by drifting, which can expand the dynamic limit of the vehicle and increase the safety of high-speed vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flow chart of an emergency obstacle avoidance path planning and control method based on model predictive control of the present invention;

[0054] Figure 2 It is a drift obstacle avoidance decision representation diagram in the present invention;

[0055] Figure 3 A schematic diagram of a hybrid path search and an optimized path search in the present invention;

[0056] Figure 4 It is a MPC framework diagram in an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be pointed out that the embodiments described below are intended to facilitate the understanding of the present invention and do not have any limiting effect on the present invention.

[0058] like Figure 1 As shown, an emergency obstacle avoidance path planning and control method based on model predictive control includes a decision-making part, a path planning part and a tracking control part, as follows:

[0059] (1) The first layer is the decision-making part, which is for sudden obstacles. When the vehicle is running at high speed, once an obstacle suddenly appears in front of it (it can be cargo, animals or people), it cannot be avoided by conventional turns. The vehicle controller needs to make a decision to drift to avoid the obstacle. The judgment that conventional turns cannot avoid obstacles is based on the current driving speed of the vehicle and the critical condition of rollover. The turning radius based on the critical condition of rollover at the current speed is called the minimum turning radius. If the turning radius cannot complete the obstacle avoidance, it is necessary to avoid the obstacle by drifting.

[0060] The analysis of vehicle rollover conditions requires the use of a four-wheel vehicle model, because the critical condition for vehicle rollover is formed when one wheel is under force and the other wheel is not under force. Assuming that the left wheel of the vehicle is not under force, the right wheel will bear the gravity of the vehicle. According to the direction of vehicle travel, that is, the moment balance of the x-axis, it can be obtained:

[0061]

[0062] m is the mass of the vehicle, g is the acceleration of gravity, L is the distance between the two front wheels or the two rear wheels, f is the friction force on the wheel on the side of the vehicle when the vehicle is in the critical condition of rollover, and d is the height of the center of mass of the vehicle from the ground. v is the current vehicle speed, R is the vehicle turning radius, that is, the minimum turning radius of the vehicle under the critical condition of rollover. According to the minimum turning radius, the front wheel turning angle of the vehicle under the critical condition of rollover can be calculated as L′ is the distance between the front and rear wheels of the vehicle. When the vehicle determines that the minimum turning radius according to the rollover critical condition cannot effectively avoid obstacles based on the current speed, it needs to drift to avoid obstacles.

[0063] (2) The second layer is the path planning part. After obtaining the instruction of using drift to avoid obstacles from the first layer, the path planning part searches for a forward drifting obstacle avoidance path according to the current vehicle speed. The drifting path search is based on a hybrid path search and an optimized path search method. The front wheel angle is used to perform a dynamics-based forward path search. For each front wheel angle search point, a search weight calculation based on the optimization problem is performed. Combined with the weight of each search point, the search point with the largest weight is selected as the search point for forward path planning. After optimizing the forward output of each point, the system model state and input value at each moment are calculated as a reference for the system state and input of the next layer controller at each moment.

[0064] (3) The third layer is the tracking control part, which needs to obtain the value of the drift obstacle avoidance balance point and the reference value of the system model state and input at each moment from the second layer path planning part. After obtaining the above information, the third layer tracking controller needs to track the planned drift obstacle avoidance trajectory, use the model predictive control method for tracking control, establish the MPC framework, and construct the MPC loss function to solve the linear quadratic programming problem to solve the input at the current moment. After the solution is completed, return to the second layer to perform rolling optimization with the current state and input.

[0065] like Figure 2 As shown, when an obstacle appears, the drift obstacle avoidance decision is based on the position of the obstacle and the current speed of the vehicle as the judgment criteria. When the minimum turning radius at the current speed cannot effectively avoid the obstacle, a decision to use drifting to avoid the obstacle needs to be made.

[0066] like Figure 3 As shown in the figure, after making the decision to drift and avoid obstacles, the second-level path planning part needs to plan a drift path that not only needs to meet the requirements of obstacle avoidance, but also needs to meet the vehicle drift dynamics. The drift dynamics formula is as follows:

[0067]

[0068]

[0069]

[0070] make The equation at the drift equilibrium point can be obtained, where is the derivative of velocity, is the derivative of the vehicle sideslip angle, is the derivative of the vehicle's yaw rate, V is the vehicle speed, β is the vehicle's sideslip angle, r is the vehicle's yaw rate, δ is the front wheel turning angle, a and b are the distances from the vehicle's center of mass to the front and rear wheels, respectively, and F yf and F yr are the lateral forces on the front and rear wheels of the vehicle, respectively, and F xr is the rear wheel driving force of the vehicle, m is the vehicle mass, I z is the moment of inertia of the vehicle around the Z axis. The equation at the equilibrium point is as follows:

[0071]

[0072]

[0073]

[0074] According to the magic tire formula Since the rear wheel is in a saturated state during drifting, the lateral force of the rear wheel In the formula, and are the components of gravity acting on the front and rear wheels, respectively, and a and b are the distances from the center of mass of the vehicle to the front and rear wheels, respectively. Since there are three sets of equations and five unknown quantities in the drift dynamics equation, we use the current speed as the speed of the equilibrium point and the front wheel turning angle as the search; we determine V eq is the current speed, and the front wheel steering angle δ = -10°:-10°:-30° (counterclockwise drift) or δ = 10°:10°:30° (clockwise drift) is used as the search interval, and the value V of each item at the equilibrium point can be obtained. eq ,β eq ,r eq ,δ eq , According to the value of the equilibrium point, forward simulation can be performed to calculate the path point information at each moment. The specific calculation is as follows:

[0075]

[0076] st

[0077] Δξ(k+1)=AΔξ(k)+BΔu(k)

[0078] Δu min ≤Δu(k)≤Δu max

[0079] r k represents the predicted yaw rate at time k, r eq represents the expected value of the yaw rate when the vehicle drifts. Similarly, β k and V k They represent the predicted value of the vehicle sideslip angle and the predicted value of the vehicle body speed at time k respectively; β eq and V eq is the value of the sideslip angle and vehicle speed during steady-state drift; δ k and Fxr k are the predicted values ​​of the front wheel steering angle and rear wheel driving force at time k, δ eq and are the values ​​of the front wheel steering angle and rear wheel driving force at the equilibrium point respectively; Δx(k) and Δu(k) are the deviations of the system state quantity and the input quantity relative to the equilibrium point value at time k respectively, and the system state quantity is ξ=[r,β,V] T , the system input is u=[δ,F xr ] T .

[0080] After solving the optimization problem, calculate its path:

[0081]

[0082]

[0083]

[0084]

[0085] For each front wheel turning angle δ, there is a corresponding forward search path [(X1, Y1), (X2, Y2), …, (X n ,Y n )]. Taking counterclockwise drift as an example, the front wheel angle search interval and interval are δ=-10:-10:-30. For each front wheel angle corresponding to the forward search path, the weight corresponding to the current front wheel angle value is calculated. The weight can be expressed by the following formula:

[0086]

[0087] (X obs ,Y obs ) is the obstacle coordinate; according to the weight value H corresponding to each front wheel search angle, the largest value is selected as the drift reference path.

[0088] like Figure 4 As shown, for Figure 3 We need to control the vehicle to track the path planned by the second-layer path planner. The tracking controller framework adopts the model predictive control framework. First, the model in the framework selects the linear discrete model of dynamics, which is obtained by linearizing and discretizing the dynamic equations:

[0089] Δx(k+1)=AΔx(k)+BΔu(k)

[0090] In the formula, x(k) is the system state, u(k) is the system input, the linearization method uses the Taylor expansion method, and the discretization method uses the forward Euler method. In order to track the planned path, the state x=[e,Δφ,V,β,r] is selected T , input u=[δ,F xr ] T , e is the lateral error of the vehicle's current position from the reference path, Δφ is the deviation between the current heading angle and the desired heading angle, V is the vehicle's center of mass velocity, β is the center of mass sideslip angle, and r is the vehicle's yaw rate. The input δ is the front wheel steering angle, F xr is the rear wheel driving force. The system matrix A and input matrix B are obtained by

[0091]

[0092]

[0093] In the above matrix y1,y2,…,y n (n=5) is the following equation:

[0094]

[0095]

[0096]

[0097]

[0098]

[0099] After the system model is proposed, it is necessary to construct the MPC optimization problem and solve the optimization problem in the form of solving the QP problem. The MPC optimization problem requires optimizing the model input at the time (k~k+Np-1), and the optimization function is the following expression:

[0100]

[0101] st

[0102] Δx(k+1)=AΔx(k)+BΔu(k)

[0103] Δu min ≤Δu(k)≤Δu max

[0104] x(k) is the input of the model at time k x = [e, Δφ, V, β, r] T ,x eq is the value of the equilibrium point of the drift process model state quantity, u(k) is the model input quantity at time k u=[δ,F xr ] T The value of Δu min and Δu max are the minimum and maximum allowable input deviations at time k and time k-1, and A and B are the system matrix and input matrix obtained after linearization + discretization, respectively.

[0105] The input obtained by the optimization function will be input into the model. After running the input deviation at time k, secondary path planning and tracking will be performed on the current vehicle state and input. Until the vehicle successfully avoids obstacles using drift, the planning will be terminated and the vehicle will return to the initial path to continue walking.

[0106] The embodiments described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An emergency obstacle avoidance path planning and control method based on model predictive control, characterized in that: It includes decision-making part, path planning part and tracking control part, as follows: Decision-making part: When the vehicle detects an obstacle during high-speed operation and determines that it cannot avoid the obstacle through conventional turns, it makes a decision to drift to avoid the obstacle; Path planning part: After the decision to drift and avoid obstacles is made, a forward drift and obstacle avoidance path is searched according to the current vehicle speed to generate a drift and obstacle avoidance path; the forward drift and obstacle avoidance path search is based on a hybrid path search and an optimized path search method; the hybrid path search and optimized path search method uses the front wheel angle to search for a forward path based on vehicle dynamics; for each front wheel angle search point, a search weight calculation based on an optimization problem is performed, and the search point with the largest weight is selected as the search point for forward path planning in combination with the weight of each search point, and the forward output of each point is optimized to calculate the system model state and input value at each moment, which are used as a reference for the system model state and input of the model predictive controller at each moment in the tracking control part; Tracking control part: Based on the linearized vehicle dynamics model, a model predictive controller with the equilibrium point during drift as the state quantity is used to track the generated drift obstacle avoidance path; the model predictive controller adopts the model predictive control framework, and the model in the model predictive control framework adopts a linear discrete model of dynamics, which is obtained by linearizing and discretizing the dynamic equations: Δx(k+1)=AΔx(k)+BΔu(k) In the formula, x(k) is the system state, u(k) is the system input, the linearization method uses the Taylor expansion method, and the discretization method uses the forward Euler method; in order to track the planned path, the state x = [e, Δφ, V, β, r] is selected T , input u=[δ,F xr ] T , e is the lateral error of the vehicle's current position from the reference path, Δφ is the deviation between the current heading angle and the desired heading angle, V is the vehicle's center of mass velocity, β is the center of mass sideslip angle, and r is the vehicle's yaw rate; the input δ is the front wheel steering angle, F xr is the rear wheel driving force; the system matrix A and input matrix B are obtained by In the above matrix, m = 2, n = 5, y1, y2, ..., y n The following equation: a and b are the distances from the center of mass of the vehicle to the front and rear wheels, respectively. yf and F yr are the lateral forces on the front and rear wheels of the vehicle respectively; m is the vehicle mass, I z is the moment of inertia of the vehicle around the Z axis; After the system model is proposed, it is necessary to construct the optimization problem of the model predictive controller and solve the optimization problem in the form of solving the QP problem. The model predictive controller optimization problem needs to optimize the model input at the time (k~k+Np-1). The optimization function is the following expression: st Δx(k+1)=AΔx(k)+BΔu(k) Δu mib ≤Δu(k)≤Δu max x(k) is the input of the model at time k x = [e, Δφ, V, β, r] T ,x eq is the value of the equilibrium point of the drift process model state quantity, u(k) is the model input quantity at time k u=[δ,F xr ] T The value of Δu min and Δu max are the minimum and maximum allowable input deviations at time k and time k-1, respectively. A and B are the system matrix and input matrix obtained after linearization and discretization of the vehicle dynamics model of the tracking control process.

2. The method for emergency obstacle avoidance path planning and control based on model predictive control according to claim 1, characterized in that: In the decision-making part, after detecting an obstacle, the vehicle determines whether the minimum turning radius can be used to avoid the obstacle based on the obstacle position and the current vehicle speed. If it is determined that the minimum turning radius cannot effectively avoid the obstacle, a decision is made to use drifting to avoid the obstacle.

3. The method for emergency obstacle avoidance path planning and control based on model predictive control according to claim 1, characterized in that: The front wheel turning angle is used as the search point. When the vehicle drifts counterclockwise, the front wheel turning angle takes a value of -10°:-10°:-30°. When the vehicle drifts clockwise, the front wheel turning angle takes a value of 10°:10°:30°.

4. The method for emergency obstacle avoidance path planning and control based on model predictive control according to claim 1, characterized in that: The generated drift obstacle avoidance path satisfies both obstacle avoidance requirements and vehicle drift dynamics. The formula for vehicle drift dynamics is as follows: in, is the derivative of velocity, is the derivative of the vehicle sideslip angle, is the derivative of the vehicle's yaw rate, V is the vehicle speed, β is the vehicle's sideslip angle, r is the vehicle's yaw rate, δ is the front wheel turning angle, a and b are the distances from the vehicle's center of mass to the front and rear wheels, respectively, and F yf and F yr are the lateral forces on the front and rear wheels of the vehicle, respectively, and F xr is the rear wheel driving force of the vehicle, m is the vehicle mass, I z is the moment of inertia of the vehicle around the Z axis; The equation at the drift equilibrium point is obtained, and the values ​​of various parameters at the drift equilibrium point are obtained after solving V eq ,β eq ,r eq ,δ eq , Among them, β eq and V eq is the value of the sideslip angle and vehicle speed during steady-state drift, r eq represents the expected value of the yaw rate when the vehicle drifts, δ eq and are the values ​​of the front wheel steering angle and rear wheel driving force at the equilibrium point respectively.

5. The method for emergency obstacle avoidance path planning and control based on model predictive control according to claim 4, characterized in that: According to the value of the vehicle drift balance point, a heuristic search function is used to forward fit the drift obstacle avoidance path. The heuristic search function takes the form of the following optimization function: st Δξ(k+1)=A′Δξ(k)+B′Δu(k) Δu min ≤Δu(k)≤Δu max In the formula, r k represents the predicted yaw rate at time k, r eq represents the expected value of the yaw rate when the vehicle drifts; similarly, β k and V k They represent the predicted value of the vehicle sideslip angle and the predicted value of the vehicle body speed at time k respectively; β eq and V eq is the value of the sideslip angle and vehicle speed during steady-state drift; δ k and Fxr k are the predicted values ​​of the front wheel steering angle and rear wheel driving force at time k, δ eq and are the values ​​of the front wheel steering angle and rear wheel driving force at the equilibrium point respectively; Δξ(k) and Δu(k) are the deviations of the system state quantity and the input quantity relative to the equilibrium point value at time k respectively, and the system state quantity is ξ=[r,β,V] T , the system input is u=[δ,F xr ] T ; A′ and B′ are the system matrix and input matrix obtained after linearization and discretization of the vehicle drift dynamics model.

6. The method for emergency obstacle avoidance path planning and control based on model predictive control according to claim 5, characterized in that: After solving the optimization problem, the drift path is calculated. The formula is as follows: For each front wheel turning angle δ, there is a corresponding forward search path [(X1, Y1), (X2, Y2), …, (X n ,Y n )]; For each front wheel turning angle corresponding to the forward search path, the weight corresponding to the current front wheel turning angle value is calculated, and the weight is expressed by the following formula: (X obs ,Y obs ) is the obstacle coordinate; according to the weight value H corresponding to each front wheel search angle, the largest value is selected as the drift obstacle avoidance path.

Images