Automatic driving automobile collision avoidance method based on fuzzy model predictive control

Abstract

The invention relates to an automatic driving automobile collision avoidance method based on fuzzy model predictive control, and belongs to the technical field of collision avoidance of automatic driving automobiles. Comprising the following steps: constructing a transverse dynamic model of the autonomous vehicle; designing a fuzzy rule according to the T-S fuzzy logic, converting the transverse dynamic model into a fuzzy dynamic model based on the fuzzy rule, and discretizing the fuzzy dynamic model; designing collision avoidance constraints of the fuzzy dynamic model based on the control quantity and the motion state; and constructing a fuzzy model prediction controller based on the discretized fuzzy dynamic model and the collision avoidance constraint, performing optimization solution on the fuzzy model prediction controller to obtain a control quantity, and controlling the automatic driving vehicle to perform collision avoidance based on the control quantity.

Description

Technical Field

[0001] The present invention belongs to the technical field of collision avoidance of an autonomous driving vehicle, and in particular relates to a collision avoidance method for an autonomous driving vehicle based on fuzzy model predictive control. Background Art

[0002] With the rapid development of autonomous driving technology, ensuring that autonomous vehicles can safely and efficiently avoid collisions in various complex road environments has become a crucial technical challenge. In recent years, the development of autonomous vehicles has greatly improved the safety and efficiency of road traffic. Driving safety is an issue that cannot be ignored in the development process. Due to the existence of obstacles on the road, how to ensure that the vehicle can effectively perceive, judge and avoid possible collisions in various complex road environments is an important part of autonomous driving vehicle research. Autonomous driving technology is developing towards the goal of optimizing control algorithms, eliminating control delays, and ensuring that vehicles can efficiently and safely avoid collisions. In addition, the comfort and stability of the vehicle must be considered to avoid excessive acceleration or deceleration during collision avoidance, which may interfere with the passengers' riding experience.

[0003] At present, many research works are devoted to the field of collision avoidance control of autonomous vehicles, mainly focusing on improving the vehicle's ability to perceive the surrounding environment, accurately planning obstacle avoidance paths, and achieving efficient vehicle control. Autonomous vehicles are equipped with a variety of sensors, such as lidar, radar, cameras, and ultrasonic sensors, to achieve all-round perception of the surrounding environment. These sensors can capture road information, vehicle position, pedestrian dynamics, and other potential obstacles in real time, providing accurate environmental data for the vehicle. Autonomous vehicles use advanced algorithms for processing and analysis to build a detailed environmental model. Based on this model, the vehicle can identify the drivable area and obstacle locations, and plan the optimal driving path accordingly. In particular, when encountering obstacles, the vehicle will quickly generate an obstacle avoidance trajectory to ensure that it can rejoin the preset nominal trajectory after avoiding the obstacle, which ensures both safety and smooth driving. When planning the obstacle avoidance trajectory, autonomous vehicles will use methods such as parameterized S-shaped function curves to ensure the smoothness and feasibility of the trajectory.

[0004] Design trajectory tracking control methods, including control algorithms based on road geometry principles, such as pure tracking control, Stanley control, Alice control, etc.; path tracking control algorithms based on classical control theory, such as PID control, linear feedback control, etc. And path tracking control algorithms based on modern control theory, such as MPC control, LQR control, etc. At the same time, self-driving cars will use a lateral controller with feedforward plus robust feedback to achieve accurate tracking of the planned trajectory. This controller can calculate appropriate control instructions such as steering angle, acceleration and braking force based on the current state of the vehicle and the target trajectory, ensuring that the vehicle can accurately and stably drive along the planned trajectory.

[0005] The application of model predictive control in autonomous driving and collision avoidance control is reflected in its powerful prediction and optimization capabilities. In the field of autonomous driving, model predictive control predicts the future motion state of the vehicle by constructing a dynamic model of the vehicle, and optimizes the control strategy in combination with real-time feedback information to achieve accurate trajectory tracking and path planning. In collision avoidance control, model predictive control can predict the relative position and speed relationship between the vehicle and the obstacle, calculate the optimal collision avoidance trajectory in advance, and effectively avoid potential collisions by controlling the vehicle's speed, direction and other parameters. This control method not only improves the safety of autonomous driving vehicles, but also ensures comfort and stability during driving, providing strong support for the development of autonomous driving technology.

[0006] Most existing path planning algorithms for obstacle avoidance control technologies for autonomous vehicles may be computationally intensive and inefficient in complex environments, and the generated paths may not be optimal. At the same time, trajectory tracking control methods may perform poorly when faced with nonlinear and time-varying systems, and require precise parameter adjustments. Although model predictive control predicts future vehicle states to avoid possible collisions, it has high computational complexity and requires high accuracy of the model. These problems may affect the obstacle avoidance performance and safety of autonomous vehicles. Summary of the invention

[0007] The technical problems to be solved by the present invention are:

[0008] In order to avoid the shortcomings of the prior art, the present invention provides an automatic driving vehicle collision avoidance method based on fuzzy model predictive control, which is used to solve the problem that the existing model predictive control is complex in calculation and causes errors.

[0009] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0010] A collision avoidance method for an automatic driving vehicle based on fuzzy model predictive control, characterized by comprising:

[0011] S1: Construct a lateral dynamics model of an autonomous vehicle;

[0012] S2: Design fuzzy rules according to TS fuzzy logic, transform the lateral dynamics model into a fuzzy dynamics model based on the fuzzy rules and discretize it;

[0013] S3: Design collision avoidance constraints of fuzzy dynamics model based on control variables and motion states;

[0014] S4: A fuzzy model predictive controller is constructed based on the discretized fuzzy dynamic model and collision avoidance constraints. The fuzzy model predictive controller is optimized and solved to obtain the control quantity, and the autonomous driving vehicle is controlled to avoid collision based on the control quantity.

[0015] A further technical solution of the present invention: The lateral dynamics model of the autonomous driving vehicle is:

[0016]

[0017] Where x = [y e ,ψ e ,v y ,ω z ] T is the state vector of the vehicle system, y e is the current lateral position deviation of the vehicle, ψ e is the vehicle yaw rate error, v y is the lateral velocity of the vehicle, ω z is the yaw rate of the vehicle; u is the control variable of the vehicle model, which represents the front wheel steering angle, that is, u = δ; d is the vehicle model error, expressed as d = [d y ,d ψ ,d v ,d ω ] T , d y ,d ψ ,d v ,d ω are the errors of the four state vectors respectively; B d is the model error matrix, C is the output matrix; A is the system matrix, B is the control matrix, expressed as:

[0018]

[0019] Among them, v x is the vehicle longitudinal velocity, C f and C r are the turning stiffness of the front and rear wheels respectively, l f and l r Represent the distance between the center of mass and the front and rear wheels, m and I respectively z are the mass and moment of inertia of the vehicle respectively.

[0020] A further technical solution of the present invention is as follows: the discretized fuzzy dynamics model is specifically:

[0021] x k+1 =(I+TA k )x k +TB k u k +TB d,k d k

[0022] Where T is the sampling step of the system, A k , B k and B d,k is the system matrix after fuzzy processing.

[0023] A further technical solution of the present invention is: the collision avoidance constraint includes an input constraint and a state constraint,

[0024] The input constraints include input and input increment constraints, specifically:

[0025] u min <u k <u max

[0026] Δu min <Δu k <Δu max

[0027] Among them, u min and u max are the lower and upper limits of the control input, Δu k Refers to the control increment at the kth step, and its upper and lower bounds are Δu min and Δu max ;

[0028] The state constraints include position and speed constraints, specifically:

[0029] v y,min <v y,k <v y,max

[0030] ψ min <ψ k <ψ max

[0031] ω min <ω k <ω max

[0032] in,(·) min and(·) max represent the upper and lower bounds of the constraint variables respectively.

[0033] A further technical solution of the present invention is as follows: the control quantity is obtained by optimizing and solving the fuzzy model predictive controller, specifically:

[0034] Design the objective function of the fuzzy model predictive controller:

[0035]

[0036] Among them, N p and N c are the prediction step size and control step size respectively, Q, R and H are the weight matrices of the objective function;

[0037] The objective function is transformed into an optimization problem:

[0038]

[0039] A computer system, characterized in that it comprises: one or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned method.

[0040] A computer-readable storage medium, characterized in that it stores computer-executable instructions, which are used to implement the above method when executed.

[0041] A computer program product, characterized by comprising computer executable instructions, wherein the instructions are used to implement the above method when executed.

[0042] The beneficial effects of the present invention are:

[0043] The present invention provides a collision avoidance method for an autonomous driving vehicle based on fuzzy model predictive control, so as to achieve effective obstacle avoidance of the autonomous driving vehicle facing multiple obstacles in a dual-lane change driving scenario. A vehicle lateral dynamics model is established, and the vehicle tire stiffness and moment of inertia are taken into account. Fuzzy rules are formulated and a fuzzy logic controller is designed to blur the nonlinear vehicle dynamics model into a weighted form of a linear model. Based on the fuzzy vehicle dynamics model, a fuzzy model predictive controller is designed, taking into account the state and control quantity constraints during the vehicle motion process. Ultimately, while ensuring the stability of the vehicle motion, effective collision avoidance of the autonomous driving vehicle is achieved with minimal computational effort. The details are as follows:

[0044] 1. The present invention establishes a fuzzy vehicle dynamics model, including the establishment of a vehicle dynamics model and a fuzzy controller. By setting fuzzy logic rules, the nonlinear vehicle model is effectively simplified. The fuzzy vehicle model is used to simplify the nonlinear vehicle model into a weighted sum of a linear system.

[0045] 2. By considering the constraints in the vehicle collision avoidance process, using the fuzzy vehicle model and control quantity and state quantity constraints, a model predictive controller is designed to reduce the amount of calculation and achieve effective avoidance of obstacles. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0047] Figure 1 Flow chart of the method of the present invention.

[0048] Figure 2 Vehicle motion model.

[0049] Figure 3 The trajectory of the collision avoidance vehicle.

[0050] Figure 4 Vehicle lateral speed change curve.

[0051] Figure 5 Vehicle heading angle change curve.

[0052] Figure 6 Vehicle heading angular velocity change curve. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0054] In the core field of autonomous driving technology, driving safety occupies a pivotal position. Regarding the collision avoidance problem of autonomous driving vehicles, due to the complexity of nonlinear vehicle dynamics models and the diversity of obstacles in the road environment, how to effectively avoid collisions and maintain stable operation of the vehicle in the presence of obstacles has become a challenging problem that needs to be solved urgently. The present invention deeply studies the collision avoidance control mechanism of autonomous driving vehicles and proposes an optimized fuzzy model predictive control algorithm. The present invention establishes a series of fuzzy rules based on the Takagi–Sugeno fuzzy method to transform the originally complex nonlinear vehicle system into a weighted sum form of a series of linear affine models. This transformation not only significantly reduces the computational cost of the algorithm, but also ensures the maintenance of control accuracy, thereby ensuring the collision avoidance performance of autonomous driving vehicles in complex road environments.

[0055] The present invention provides an autonomous driving vehicle collision avoidance method based on fuzzy model predictive control, which includes constructing a vehicle dynamics model, designing fuzzy logic, and designing collision avoidance constraints. A fuzzy model predictive controller is designed based on the vehicle dynamics model, fuzzy logic, and collision avoidance constraints to control the autonomous driving vehicle to avoid collision.

[0056] Building a vehicle dynamics model

[0057] First, a lateral dynamics model of an autonomous vehicle is designed to better evaluate the performance of the vehicle dynamic system and achieve precise control of the autonomous vehicle. In order to ensure the accuracy of the model and simplify the subsequent fuzzy model predictive controller design, the present invention establishes a lateral dynamics model of the vehicle, such as Figure 2 shown.

[0058] like Figure 2 As shown in Figure 1, X and Y represent the global position of the vehicle in the longitudinal and lateral directions, respectively. Taking the center of mass of the vehicle as the reference point, l f and l r Respectively represent the distance between the center of mass and the front and rear wheels. And F yf and F yr are defined as the lateral tire forces acting on the front and rear wheels, respectively. δ is the front wheel steering angle.

[0059] The lateral position deviation and heading angle deviation of the vehicle are defined as

[0060]

[0061] Where y is the lateral position of the vehicle and ψ is the heading angle of the vehicle. d and ψ d are the desired lateral position and heading angle of the vehicle respectively. Then the lateral dynamics model of the vehicle can be expressed as:

[0062]

[0063] Among them, v x is the vehicle longitudinal velocity, v y is the lateral velocity of the vehicle, y e is the current lateral position deviation of the vehicle. z is the vehicle’s yaw rate, ω d is the reference yaw rate of the vehicle, then the vehicle yaw rate error can be expressed as the difference between the vehicle yaw rate and the reference yaw rate ψ e .

[0064] Assuming that the heading angle error of the autonomous vehicle is small enough, further simplification can be performed to obtain the following approximation:

[0065]

[0066] Therefore, the vehicle lateral dynamics model can be transformed into:

[0067]

[0068] In order to accurately control the vehicle acceleration, by analyzing the relationship between the vehicle tire force and the vehicle speed change rate, and considering the relationship between the control force and acceleration, the vehicle dynamics model is constructed as follows:

[0069]

[0070] Among them, m and I z are the mass and moment of inertia of the vehicle respectively.

[0071] Under normal vehicle driving conditions, in order to facilitate the design of the auxiliary controller and ensure the calculation accuracy, the linear tire model is used to calculate the vehicle lateral tire force, which can be expressed as:

[0072]

[0073] Among them, C f and C r are the turning stiffness of the front and rear wheels respectively. Based on formulas (1)-(7), the lateral motion model of the vehicle can be expressed as:

[0074]

[0075] Where x is the state vector of the vehicle system, expressed as x = [y e ,ψ e ,v y ,ω z ] T u is the control variable of the vehicle model, which represents the front wheel steering angle, that is, u = δ. A is the system matrix, and B is the control matrix. d is the vehicle model error, which can be expressed as d = [d y ,d ψ ,d v ,d ω ] T , the corresponding model error matrix can be expressed as B d The system matrix and control matrix can be expressed as:

[0076]

[0077] Designing fuzzy logic

[0078] In the actual driving process of the vehicle, due to the influence of various factors such as tire-ground friction, air resistance, vehicle load, etc., the actual performance of the lateral stiffness may show nonlinear changes. By establishing a fuzzy model of the lateral stiffness and longitudinal speed of the vehicle model, the vehicle's trajectory tracking ability and collision avoidance ability are guaranteed.

[0079] In the fuzzy logic model, the longitudinal speed and tire stiffness are usually kept within a certain range. The designed fuzzy logic model enables us to effectively predict and control these two key parameters in the presence of uncertainty and ambiguity. Due to the physical limitations of the vehicle, the TS fuzzy system is used to linearize the vehicle lateral dynamics model at different upper and lower boundaries of the variables. The range of uncertain variables in the system model is described as:

[0080]

[0081] in, * and Respectively represent the upper and lower bounds of the fuzzy variables. x ,1 / v x ,C f ,C r}.

[0082] The membership functions of different variables and their corresponding boundaries can be derived as follows:

[0083]

[0084] At the same time, in the design of the fuzzy controller, the membership function needs to meet the following conditions:

[0085]

[0086] Where j = {1, 2, 3, 4}. j represents the membership variables corresponding to the membership parameters of different variables, M j,l and M j,s is the membership function on the upper and lower boundaries of the above variables.

[0087] Based on TS fuzzy logic, the fuzzy rules shown in Table 1 are established to linearize the nonlinear model.

[0088] Table 1 Fuzzy rules

[0089]

[0090] In Table 1, s and l represent the smaller and larger values ​​of the membership parameter values ​​from the fuzzy controller input, respectively. At the same time, the membership parameter must satisfy:

[0091]

[0092] Under the weighted combination of local linear models, the overall lateral dynamics model is derived as:

[0093]

[0094] Among them, the output matrix C i and the model error matrix B d,i Defined as C i =I and B d,i =I. In addition, A i and B i are the state and input matrices of the local linear model.

[0095] According to formula (15), the discretization form of the vehicle system is derived based on the Euler method:

[0096] x k+1 =(I+TA k )x k +TB k u k +TB d,k d k (16)

[0097] Where T is the sampling step of the system, A k , B k and B d,k is the system matrix after fuzzy processing.

[0098] Design collision avoidance constraints for vehicle system models

[0099] In the present invention, obstacles are transformed into constraints in the model predictive controller. The design of the model predictive controller needs to consider state and input constraints, which is the key to ensure that the control system is both safe and efficient in practice. Input constraints are restrictions on control signals or manipulated variables to prevent inputs from exceeding the capabilities of the controlled vehicle actuators, resulting in control failures or system instability. The input and input increment constraints are defined as:

[0100] u min <u k <u max (17)

[0101] Δu min <Δu k <Δu max (18)

[0102] Among them, u min and u max are the lower and upper limits of the control input, Δu k Refers to the control increment at the kth step, and its upper and lower bounds are Δu min and Δu max To ensure that the system does not enter an unacceptable or dangerous state that could lead to a safety incident, the position and speed of the controlled vehicle are limited to:

[0103] v y,min<v y,k <v y,max (19)

[0104] ψ min <ψ k <ψ max (20)

[0105] ω min <ω k <ω max (twenty one)

[0106] in,(·) min and(·) max represent the upper and lower bounds of the constraint variables respectively.

[0107] Design of fuzzy model predictive controller

[0108] In the present invention, the multi-constrained fuzzy control problem is transformed into a model predictive control problem. Based on the fuzzy vehicle model, the collision avoidance strategy in the collision avoidance scenario is designed. In order to minimize the difference between the vehicle motion state and the reference state, while considering the energy cost and driving stability, the control force and control increment are considered in the design of the fuzzy model predictive controller. Therefore, the objective function is designed as:

[0109]

[0110] Among them, N p and N c are the prediction step size and control step size respectively, Q, R and H are the weight matrices of the objective function.

[0111] Therefore, according to the collision avoidance constraints of the vehicle system model, the fuzzy model predictive control problem of the autonomous vehicle is described as the following optimization problem:

[0112]

[0113] Simulation experiment

[0114] The proposed collision avoidance control technology is verified by simulating a dual-lane, multi-obstacle autonomous driving obstacle avoidance scenario. In the simulation study, the autonomous driving vehicle will use the traditional model predictive control and the improved fuzzy model predictive control in the present invention to avoid obstacles in the dual lanes, verifying the effectiveness of the proposed fuzzy model predictive control algorithm obstacle avoidance strategy.

[0115] In the dual-lane, multi-obstacle autonomous driving obstacle avoidance scenario, the correct obstacle avoidance strategy needs to be completed in a short time, which is a challenge for autonomous vehicles. The vehicle motion trajectory using traditional model predictive control and improved fuzzy model predictive control is as follows Figure 3 shown.

[0116] like Figure 3 As shown in the figure, in the collision avoidance of the first obstacle, both the self-driving cars using traditional model predictive control and the improved fuzzy model predictive control made collision avoidance decisions in advance and passed smoothly. However, due to the speed and distance limitations, at t = 15s, the vehicle trajectory using traditional model predictive control is about to collide with the second obstacle, while at the same time, the vehicle using fuzzy model predictive control can react to the obstacle one step earlier, avoiding a possible collision. Therefore, it is difficult for traditional fuzzy model predictive control to avoid collision with the second obstacle during driving, posing a threat to driving safety.

[0117] The design of the fuzzy controller is conducive to reducing the amount of calculation in solving the optimization problem, and at the same time makes the vehicle nominal model closer to the real model. During the movement of the autonomous vehicle, it is necessary to continuously adjust the speed and heading angle to avoid collisions during the teaching process. The lateral speed and heading angle changes of the autonomous vehicle using the traditional model predictive control and the improved fuzzy model predictive control are shown in the figure. Figure 4 and Figure 5 shown.

[0118] like Figure 4 As shown in the figure, the longitudinal speed always changes within a certain range. If the longitudinal speed is too large or too small, it will pose a threat to the driving safety of the vehicle. The lateral speed of the vehicle reaches a maximum value of 0.2m / s at around the 8th and 14th seconds, both of which occur during the vehicle's obstacle avoidance steering process around two obstacles. This is the reason for a certain amount of overshoot during the vehicle's obstacle avoidance process. Figure 5 As shown in the figure, during the dual-lane obstacle avoidance process, the maximum yaw angle is about 0.25 rad. After avoiding the first obstacle, the heading angle of the autonomous driving car using the improved fuzzy model predictive control changes less, ensuring driving stability. During the obstacle avoidance process of the second obstacle, the vehicle using the traditional model predictive control starts to change lanes at around 14 seconds, which cannot fully guarantee the vehicle's safe obstacle avoidance. The vehicle using the improved fuzzy model predictive control has a smaller overshoot and adjusts the heading angle from 0 rad to -0.23 rad faster, making the correct obstacle avoidance decision.

[0119] Precise control of the angular velocity of the vehicle is essential to ensure stable operation and precise maneuvering. The angular velocity of the autonomous vehicle changes as follows: Figure 6 shown.

[0120] like Figure 6 As shown in the figure, the vehicle's angular velocity reaches -0.03 rad / s at the 8th second. At this time, the vehicle is performing obstacle avoidance and lane change. After the first obstacle avoidance is completed, the angular velocity tends to 0 rad / s, but due to the change of state during the obstacle avoidance process, it cannot completely reach 0 rad / s.

[0121] The autonomous vehicle using fuzzy model predictive control changes with a smaller jitter at 0 rad / s, and starts the second obstacle avoidance movement within 13-14 seconds. The heading angular velocity changes from -0.03 rad / s to 0.02 rad / s.

[0122] Compared with the method of the present invention, the traditional fuzzy model predictive control algorithm is difficult to effectively avoid obstacles in complex scenarios when dealing with nonlinear systems. The fuzzy model predictive control collision avoidance technology proposed in the present invention greatly reduces the amount of calculation of the system while ensuring effective collision avoidance of obstacles, has stronger adaptability to the uncertainty and complexity of the system, and also reduces the controller's dependence on the precise model of the system, thereby improving the robustness of the system.

[0123] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should be included in the protection scope of the present invention.

Claims

1. A collision avoidance method for an autonomous driving vehicle based on fuzzy model predictive control, characterized in that: include: S1: Construct a lateral dynamics model of an autonomous vehicle; S2: Design fuzzy rules according to TS fuzzy logic, transform the lateral dynamics model into a fuzzy dynamics model based on the fuzzy rules and discretize it; S3: Design collision avoidance constraints of fuzzy dynamics model based on control variables and motion states; S4: A fuzzy model predictive controller is constructed based on the discretized fuzzy dynamic model and collision avoidance constraints. The fuzzy model predictive controller is optimized and solved to obtain the control quantity, and the autonomous driving vehicle is controlled to avoid collision based on the control quantity.

2. The method for avoiding collision of an automatic driving vehicle based on fuzzy model predictive control according to claim 1, characterized in that: The lateral dynamics model of the autonomous driving vehicle is: Where x = [y e ,ψ e ,v y ,ω z ] T is the state vector of the vehicle system, y e is the current lateral position deviation of the vehicle, ψ e is the vehicle yaw rate error, v y is the lateral velocity of the vehicle, ω z is the yaw rate of the vehicle; u is the control variable of the vehicle model, which represents the front wheel steering angle, that is, u = δ; d is the vehicle model error, expressed as d = [d y ,d ψ ,d v ,d ω ] T , d y ,d ψ ,d v ,d ω are the errors of the four state vectors respectively; B d is the model error matrix, C is the output matrix; A is the system matrix, B is the control matrix, expressed as: Among them, v x is the vehicle longitudinal velocity, C f and C r are the turning stiffness of the front and rear wheels respectively, l f and l r Represent the distance between the center of mass and the front and rear wheels, m and I respectively z are the mass and moment of inertia of the vehicle respectively.

3. The method for avoiding collision of an automatic driving vehicle based on fuzzy model predictive control according to claim 2, characterized in that: The discretized fuzzy dynamics model is specifically: x k+1 =(I+TA k )x k +TB k u k +TB d,k d k Where T is the sampling step of the system, A k , B k and B d,k is the system matrix after fuzzy processing.

4. The method for avoiding collision of an automatic driving vehicle based on fuzzy model predictive control according to claim 3, characterized in that: The collision avoidance constraints include input constraints and state constraints. The input constraints include input and input increment constraints, specifically: in min <in k <in max Δu min <Δu k <Δu max Among them, u min and u max are the lower and upper limits of the control input, Δu k Refers to the control increment at the kth step, and its upper and lower bounds are Δu min and Δu max ; The state constraints include position and speed constraints, specifically: v y,min <v y,k <v y,max ψ min <ψ k <ψ max oh min <ω k <ω max in,(·) min and(·) max represent the upper and lower bounds of the constraint variables respectively.

5. The method for collision avoidance of an automatic driving vehicle based on fuzzy model predictive control according to claim 1, characterized in that: The control quantity obtained by optimizing and solving the fuzzy model predictive controller is specifically: Design the objective function of the fuzzy model predictive controller: Among them, N p and N c are the prediction step size and control step size respectively, Q, R and H are the weight matrices of the objective function; The objective function is transformed into an optimization problem:

6. A computer system, characterized in that include: One or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method of claim 1.

7. A computer-readable storage medium, characterized in that Computer executable instructions are stored, and when the instructions are executed, they are used to implement the method of claim 1.

8. A computer program product, characterized in that The invention comprises computer executable instructions, which are used to implement the method of claim 1 when being executed.

Images