A trajectory tracking control method for electric vehicles

Abstract

A method for controlling the trajectory of an electric vehicle includes: establishing an adaptive preview model, establishing an objective function of the adaptive preview model based on a gray wolf algorithm, and solving the objective function to obtain preview point coordinate information; establishing a vehicle dynamics model, and then constructing a tire cornering stiffness correction model to obtain a corrected tire cornering stiffness; combining the preview point coordinate information and the corrected tire cornering stiffness, obtaining a discrete system through the Euler method; constructing a vehicle trajectory tracking control model, and using the obtained state quantity as an evaluation index, introducing a speed pause particle swarm algorithm to optimize the state quantity weight coefficient matrix in real time; constructing a feedforward control model based on the corrected tire cornering stiffness to obtain a feedforward front wheel angle, and obtaining a final front wheel angle based on the feedforward front wheel angle and the front wheel angle output by the vehicle trajectory tracking control model. The present invention can meet the trajectory tracking accuracy of electric vehicles when they are traveling at high speed on a large curvature road.

Description

Technical Field

[0001] The invention relates to the technical field of automobiles, and in particular to a trajectory tracking control method for an electric vehicle. Background Art

[0002] With the rapid development of autonomous driving technology, vehicle trajectory tracking control faces increasingly higher requirements for accuracy and stability. In complex driving environments, vehicles need to perceive their position, speed, and dynamic characteristics in real time in order to effectively adjust their motion trajectories. Traditional trajectory tracking methods often rely on fixed control strategies, which are difficult to adapt to dynamically changing environments and driving conditions, and therefore cannot meet the needs of high-precision driving.

[0003] When a vehicle is driving at high speed on a road with large curvature, it needs a large lateral force, which can easily cause tire slippage, directly affecting the vehicle's handling performance and control accuracy. At the same time, when driving at high speed, the dynamic characteristics of the vehicle will become more complex, and the steering response will lag, resulting in an increase in tracking error.

[0004] To this end, the present invention proposes an electric vehicle trajectory tracking control method. By introducing adaptive preview technology, the vehicle speed and preview distance are dynamically adjusted to improve the tracking accuracy. At the same time, considering the existence of steady-state error in the system, the present invention establishes a feedforward control model based on tire cornering stiffness correction to reduce the steady-state error of the system. Summary of the invention

[0005] The object of the present invention is to provide a trajectory tracking control method for an electric vehicle to meet the trajectory tracking accuracy when the electric vehicle is traveling at high speed on a road with large curvature.

[0006] A trajectory tracking control method for an electric vehicle, comprising:

[0007] Step S1, based on the collected vehicle position coordinate information, heading angle, and lateral and longitudinal speed of the electric vehicle, an adaptive preview model is established, an objective function of the adaptive preview model is established based on the Grey Wolf Algorithm, and the objective function is solved to obtain preview point coordinate information;

[0008] Step S2, based on the vehicle dynamics relationship of the electric vehicle, a three-degree-of-freedom vehicle dynamics model is established, and then a tire cornering stiffness correction model is constructed to obtain a corrected tire cornering stiffness;

[0009] Step S3, obtaining the coordinate information, lateral and longitudinal speeds, curvature, and heading angle of each reference point on the reference trajectory, and combining the preview point coordinate information obtained in step S1 and the corrected tire cornering stiffness obtained in step S2 to obtain a discrete system through the Euler method;

[0010] Step S4, constructing a vehicle trajectory tracking control model according to the discrete system obtained in step S3, and using the obtained state quantity as an evaluation index, introducing a speed pause particle swarm algorithm to perform real-time optimization on the state quantity weight coefficient matrix in the vehicle trajectory tracking control model;

[0011] Step S5, constructing a feedforward control model based on the corrected tire cornering stiffness obtained in step S2, outputting the feedforward front wheel angle through the feedforward control model, and obtaining the final front wheel angle based on the feedforward front wheel angle and the front wheel angle output by the vehicle trajectory tracking control model, and feeding the final front wheel angle back to the vehicle dynamics model to form a closed-loop control to realize trajectory tracking control of the electric vehicle.

[0012] The electric vehicle trajectory tracking control method provided by the present invention has the following beneficial effects:

[0013] (1) Considering that the algorithm control has hysteresis and the vehicle does not always maintain a constant speed during driving, the present invention establishes the objective function of the adaptive preview model based on the Gray Wolf Algorithm and solves the objective function to obtain the preview point coordinate information. The preview point can be dynamically adjusted according to the real-time vehicle state and environmental changes to improve the trajectory tracking accuracy.

[0014] (2) Considering the problems of poor stability and low precision in the traditional discretization process, the present invention combines the preview point coordinate information and the corrected tire cornering stiffness to obtain a discrete system through the Euler method, which can more accurately reflect the deviation between the vehicle and the reference trajectory, thereby improving the stability and precision of the control;

[0015] (3) The present invention introduces the speed pause particle swarm algorithm to optimize the state quantity weight coefficient matrix in the vehicle trajectory tracking control model, which can enhance the adaptability and robustness of the algorithm and adapt to complex driving conditions;

[0016] (4) Taking into account the real-time nature of the tire cornering stiffness in the trajectory tracking process when the vehicle is traveling on a high-speed, high-curvature road, the present invention constructs a feedforward control model based on the corrected tire cornering stiffness. By correcting the tire cornering stiffness in real time, the steady-state error is effectively reduced, the dynamic performance of the vehicle is improved, and the vehicle handling performance is effectively improved. At the same time, the lateral error is reduced, the tracking accuracy is improved, and the vehicle performs more stably in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic flow chart of the electric vehicle trajectory tracking control method provided by the present invention;

[0018] Figure 2 It is a comparison diagram of the lateral error between the method of the present invention and the traditional trajectory tracking control method. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] See also Figure 1 , an embodiment of the present invention provides an electric vehicle trajectory tracking control method, including steps S1~S5.

[0021] Step S1, based on the collected vehicle position coordinate information, heading angle, and lateral and longitudinal speeds of the electric vehicle, an adaptive preview model is established, an objective function of the adaptive preview model is established based on the Grey Wolf Algorithm, and the objective function is solved to obtain the preview point coordinate information.

[0022] Among them, the expression of the established adaptive preview model is:

[0023] ;

[0024] ;

[0025] ;

[0026] in, X pre is the horizontal coordinate of the preview point, Y pre is the ordinate of the preview point, X is the horizontal coordinate of the vehicle’s current position, Y is the ordinate of the vehicle’s current position, v pre is the vehicle speed at the preview point, v is the current speed of the vehicle, t is the adaptive preview time, is the vehicle heading angle;

[0027] Objective function of adaptive preview model based on gray wolf optimization algorithm J h for:

[0028] J h = J 1+ J 2;

[0029] ;

[0030] ;

[0031] in, J 1 is the first evaluation function, J 2 is the first evaluation function, is the sideslip angle of the center of mass, is the ideal center of mass sideslip angle, Y ref is the vertical coordinate of the reference point.

[0032] The objective function is optimized by the Grey Wolf Optimization Algorithm. J h Optimize by updating the position of the candidate solution, find the preview time, and thus determine the optimal preview point position, and obtain the preview point coordinate information.

[0033] Step S2, based on the vehicle dynamics relationship of the electric vehicle, a three-degree-of-freedom vehicle dynamics model is established, and then a tire cornering stiffness correction model is constructed to obtain a corrected tire cornering stiffness.

[0034] Among them, the expression of the three-degree-of-freedom vehicle dynamics model is:

[0035] ;

[0036] ;

[0037] ;

[0038] in, m is the vehicle mass, F x The driving force of the vehicle, F d is the resistance of the vehicle, F y is the lateral force generated by the tire, M is the moment caused by the lateral force, is the vehicle longitudinal acceleration, is the vehicle lateral acceleration, I z is the vehicle's moment of inertia, is the vehicle heading angular velocity;

[0039] The expression of tire cornering stiffness is:

[0040] ;

[0041] ;

[0042] in, and are the cornering stiffness of the front and rear wheels respectively, and represent the lateral forces on the front and rear wheels respectively, and Represent the sideslip angles of the front and rear wheels respectively;

[0043] The least squares method is used to estimate the lateral force of the vehicle, so as to realize the correction of the tire cornering stiffness. Considering the problems of low estimation accuracy and slow convergence speed of the least squares method, the present invention introduces the mantis algorithm into the least squares method. First, the state vector is initialized. and the covariance matrix ;

[0044] ;

[0045] The least squares recursive formula based on the forgetting factor is:

[0046] ;

[0047] ;

[0048] ;

[0049] in, For the z The covariance matrix of the step, For the z -1 step covariance matrix, For the forgetting factor, is the input vector of the system, is the system input vector under the maximum number of iterations, For the z The state vector of the step, For the z -1 step state vector, is the estimation error, is the measured output;

[0050] Further determine the optimization goal of the Mantis algorithm The expression is:

[0051] ;

[0052] in, N is the total number of samples, From the first sample to the N The sum of squared errors of the estimation of samples;

[0053] Then the Mantis algorithm is initialized, the optimization target value is calculated, and then it is iterated. In each iteration, the individual position is updated and the fitness value of each individual is obtained, and it is judged whether the stopping condition is met. If not, the iteration continues. If it is met, the individual with the highest fitness value is output as the optimized forgetting factor and used in the least squares method, thereby outputting the corrected tire cornering stiffness.

[0054] Step S3, obtaining the coordinate information, lateral and longitudinal speeds, curvature, and heading angle of each reference point on the reference trajectory, and combining the preview point coordinate information obtained in step S1 and the corrected tire cornering stiffness obtained in step S2 to obtain a discrete system through the Euler method.

[0055] Wherein, step S3 specifically includes:

[0056] Based on the preview point coordinate information, traverse all reference points to find the corresponding ( X pre , Y pre ) and use it as the matching point. In the natural coordinate system, the following expression is obtained:

[0057] ;

[0058] ;

[0059] Further we get:

[0060] ;

[0061] ;

[0062] ;

[0063] ;

[0064] ;

[0065] in, is the heading angle of the matching point, is the tangent vector in the natural coordinate system, is the normal vector in the natural coordinate system, is the unit vector of error in the natural coordinate system, X mp is the horizontal coordinate of the matching point, Y mp is the ordinate of the matching point, k mp is the curvature of the matching point, is the vehicle yaw angle, is the vehicle yaw angular velocity, is the velocity of the trajectory reference point, v y is the vehicle lateral velocity, v x is the vehicle longitudinal velocity, e d is the lateral error, is the first-order derivative of the lateral error, is the heading error, is the first-order derivative of heading error;

[0066] Then the error linear differential equation in the natural coordinate system is obtained as follows:

[0067] ;

[0068] in, is the system error derivative matrix, A is the error weight matrix, B is the control weight matrix, is the system error matrix, is the control quantity matrix;

[0069] ;

[0070] ;

[0071] ;

[0072] ;

[0073] in, is the second-order derivative of the lateral error, is the second-order derivative of heading error, and The corrected cornering stiffness of the front and rear wheels, is the distance from the vehicle's center of mass to the front axle, is the distance from the vehicle's center of mass to the rear axle;

[0074] Then the discretized error equation is obtained by Euler method, which is expressed as:

[0075] ;

[0076] in, For the k +1The system error matrix at discrete time, is the discretized system error weight matrix, is the discretized control weight matrix, For the k The system error matrix at discrete time, For the k Control quantity at discrete moments;

[0077] The expression of the discrete system is further obtained as:

[0078] ;

[0079] ;

[0080] in, For the k +1 state quantity at discrete moment, For the k The state quantity at discrete moments, For the k The output at discrete moments, is the discretized output weight matrix, is the discretized state weight matrix.

[0081] Step S4, constructing a vehicle trajectory tracking control model according to the discrete system obtained in step S3, and using the obtained state quantity as an evaluation index, introducing a speed pause particle swarm algorithm to perform real-time optimization on the state quantity weight coefficient matrix in the vehicle trajectory tracking control model.

[0082] Wherein, step S4 specifically includes:

[0083] According to the discrete system obtained in step S3, the state quantity, output quantity and control quantity in the prediction time domain are obtained, and the expression is:

[0084] ;

[0085] ;

[0086] ;

[0087] in, N p For the prediction time domain, N c To control the time domain, is the state matrix in the prediction time domain, , , Respectively represent k +1 discrete moment, k +2 discrete moments, k + N p The state quantity at discrete moments, is the output matrix in the prediction time domain, , , Respectively represent k +1 discrete moment, k +2 discrete moments, k + N p The output at discrete moments, is the control quantity matrix in the prediction time domain, , , Respectively represent k +1 discrete moment, k +2 discrete moments, k + N c -1 control amount at discrete time, T represents transpose;

[0088] The vehicle trajectory tracking control model is constructed based on the state quantity, output quantity and control quantity in the prediction time domain, where the cost function J The expression is:

[0089] ;

[0090] in, Q and R are the state quantity weight coefficient matrix and the control quantity weight matrix respectively, is the relaxation factor, is the relaxation factor weight coefficient;

[0091] The optimal control problem is transformed into a standard quadratic programming problem, and the vehicle trajectory tracking control model is solved using a quadratic programming solver to obtain the front wheel steering angle and state quantity output by the vehicle trajectory tracking control model;

[0092] The obtained state quantity is used as the evaluation index, and the speed pause particle swarm algorithm is introduced to the state quantity weight coefficient matrix Q Perform real-time optimization, including:

[0093] Define the evaluation index function of the speed pause particle swarm algorithm for:

[0094] ;

[0095] in, and is the weight coefficient of the evaluation index function. The larger it is, the greater the vehicle tracking error is.

[0096] Define the objective function of the velocity pause particle swarm algorithm for:

[0097] ;

[0098] in, yes Q The norm of is the regularization parameter, For the k Evaluation index function at discrete moments;

[0099] For each initialized particle, the vehicle trajectory tracking control model is used to simulate and obtain the objective function value, which is fed back to the particle swarm. The particle speed and position are updated according to the objective function value, and the speed and position are adjusted in real time. Q The speed pause mechanism is introduced to pause the particle speed update under preset conditions, which can prevent the particles from converging to the local optimum too early.

[0100] The speed-pause particle swarm algorithm is used to calculate the state quantity weight coefficient matrix Q After optimization, the vehicle trajectory tracking control model can adjust the weighted coefficient of the state error in real time to ensure that the vehicle can accurately track the reference trajectory on high-speed and large-curvature roads.

[0101] Step S5, constructing a feedforward control model based on the corrected tire cornering stiffness obtained in step S2, outputting the feedforward front wheel angle through the feedforward control model, and obtaining the final front wheel angle based on the feedforward front wheel angle and the front wheel angle output by the vehicle trajectory tracking control model, and feeding the final front wheel angle back to the vehicle dynamics model to form a closed-loop control to realize trajectory tracking control of the electric vehicle.

[0102] Among them, since there is a system steady-state error due to the change of road curvature, a feedforward control model is constructed based on the corrected tire cornering stiffness obtained in step S2. The expression of the feedforward control model is:

[0103] ;

[0104] in, is the feedforward front wheel angle, R w is the tire radius, is the vehicle wheelbase, K is the state feedback matrix of the vehicle trajectory tracking control model.

[0105] The obtained feedforward front wheel angle The sum of the front wheel steering angle output by the vehicle trajectory tracking control model in step S4 is fed back to the vehicle dynamics model to form a closed-loop control to achieve electric vehicle trajectory tracking control. The expression is:

[0106] ;

[0107] in, The front wheel steering angle output by the vehicle trajectory tracking control model.

[0108] The present invention is simulated and tested as follows: Figure 2 The lateral error comparison diagram of the method of the present invention and the traditional trajectory tracking control method is shown in FIG. Figure 2 It can be seen that the method of the present invention greatly reduces the lateral error compared with the traditional method, and the error is 0.05m. It can be seen that the electric vehicle trajectory tracking control method proposed by the present invention can stabilize the error at a smaller value in the lateral direction, and can improve the trajectory tracking control accuracy.

[0109] In summary, the electric vehicle trajectory tracking control method provided by the present invention has the following beneficial effects:

[0110] (1) Considering that the algorithm control has hysteresis and the vehicle does not always maintain a constant speed during driving, the present invention establishes the objective function of the adaptive preview model based on the Gray Wolf Algorithm and solves the objective function to obtain the preview point coordinate information. The preview point can be dynamically adjusted according to the real-time vehicle state and environmental changes to improve the trajectory tracking accuracy.

[0111] (2) Considering the problems of poor stability and low precision in the traditional discretization process, the present invention combines the preview point coordinate information and the corrected tire cornering stiffness to obtain a discrete system through the Euler method, which can more accurately reflect the deviation between the vehicle and the reference trajectory, thereby improving the stability and precision of the control;

[0112] (3) The present invention introduces the speed pause particle swarm algorithm to optimize the state quantity weight coefficient matrix in the vehicle trajectory tracking control model, which can enhance the adaptability and robustness of the algorithm and adapt to complex driving conditions;

[0113] (4) Taking into account the real-time nature of the tire cornering stiffness in the trajectory tracking process when the vehicle is traveling on a high-speed, high-curvature road, the present invention constructs a feedforward control model based on the corrected tire cornering stiffness. By correcting the tire cornering stiffness in real time, the steady-state error is effectively reduced, the dynamic performance of the vehicle is improved, and the vehicle handling performance is effectively improved. At the same time, the lateral error is reduced, the tracking accuracy is improved, and the vehicle performs more stably in complex environments.

[0114] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0115] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A trajectory tracking control method for an electric vehicle, characterized in that: include: Step S1, based on the collected vehicle position coordinate information, heading angle, and lateral and longitudinal speed of the electric vehicle, an adaptive preview model is established, an objective function of the adaptive preview model is established based on the Grey Wolf Algorithm, and the objective function is solved to obtain preview point coordinate information; Step S2, based on the vehicle dynamics relationship of the electric vehicle, a three-degree-of-freedom vehicle dynamics model is established, and then a tire cornering stiffness correction model is constructed to obtain a corrected tire cornering stiffness; Step S3, obtaining the coordinate information, lateral and longitudinal speeds, curvature, and heading angle of each reference point on the reference trajectory, and combining the preview point coordinate information obtained in step S1 and the corrected tire cornering stiffness obtained in step S2 to obtain a discrete system through the Euler method; Step S4, constructing a vehicle trajectory tracking control model according to the discrete system obtained in step S3, and using the obtained state quantity as an evaluation index, introducing a speed pause particle swarm algorithm to perform real-time optimization on the state quantity weight coefficient matrix in the vehicle trajectory tracking control model; Step S5, constructing a feedforward control model based on the corrected tire cornering stiffness obtained in step S2, outputting the feedforward front wheel angle through the feedforward control model, and obtaining the final front wheel angle based on the feedforward front wheel angle and the front wheel angle output by the vehicle trajectory tracking control model, and feeding the final front wheel angle back to the vehicle dynamics model to form a closed-loop control to realize trajectory tracking control of the electric vehicle.

2. The electric vehicle trajectory tracking control method according to claim 1, characterized in that: In step S1, the expression of the adaptive preview model established is: ； ； ； in, X pre is the horizontal coordinate of the preview point, Y pre is the ordinate of the preview point, X is the horizontal coordinate of the vehicle’s current position, Y is the ordinate of the vehicle’s current position, v pre is the vehicle speed at the preview point, v is the current speed of the vehicle, t is the adaptive preview time, is the vehicle heading angle; Objective function of adaptive preview model based on gray wolf optimization algorithm J h for: J h = J 1+ J 2； ； ； in, J 1 is the first evaluation function, J 2 is the second evaluation function, is the sideslip angle of the center of mass, is the ideal center of mass sideslip angle, Y ref is the vertical coordinate of the reference point.

3. The electric vehicle trajectory tracking control method according to claim 2, characterized in that: In step S2, the expression of the three-degree-of-freedom vehicle dynamics model is: ； ； ； in, m is the vehicle mass, F x The driving force of the vehicle, F d is the resistance of the vehicle, F y is the lateral force generated by the tire, M is the moment caused by the lateral force, is the vehicle longitudinal acceleration, is the vehicle lateral acceleration, I z is the vehicle's moment of inertia, is the vehicle heading angular velocity; The expression of tire cornering stiffness is: ； ； in, and are the cornering stiffness of the front and rear wheels respectively, and represent the lateral forces on the front and rear wheels respectively, and Represent the sideslip angles of the front and rear wheels respectively; The least squares method is used to estimate the vehicle lateral force, so as to correct the tire cornering stiffness. The mantis algorithm is introduced into the least squares method. First, the state vector is initialized. and the covariance matrix ; ； The least squares recursive formula based on the forgetting factor is: ； ； ； in, For the z The covariance matrix of the step, For the z -1 step covariance matrix, For the forgetting factor, is the input vector of the system, is the system input vector under the maximum number of iterations, For the z The state vector of the step, For the z -1 step state vector, is the estimation error, is the measured output; Further determine the optimization goal of the Mantis algorithm The expression is: ； in, N is the total number of samples, From the first sample to the N The sum of squared errors of the estimation of samples; Then the Mantis algorithm is initialized, the optimization target value is calculated, and then it is iterated. In each iteration, the individual position is updated and the fitness value of each individual is obtained, and it is judged whether the stopping condition is met. If it is met, the individual with the highest fitness value is output as the optimized forgetting factor and used in the least squares method, thereby outputting the corrected tire cornering stiffness.

4. The electric vehicle trajectory tracking control method according to claim 3, characterized in that: Step S3 specifically includes: Based on the preview point coordinate information, traverse all reference points to find the corresponding ( X pre , Y pre ) and use it as the matching point. In the natural coordinate system, the following expression is obtained: ； ； Further we get: ； ； ； ； ； in, is the heading angle of the matching point, is the tangent vector in the natural coordinate system, is the normal vector in the natural coordinate system, is the unit vector of error in the natural coordinate system, X mp is the horizontal coordinate of the matching point, Y mp is the ordinate of the matching point, k mp is the curvature of the matching point, is the vehicle yaw angle, is the vehicle yaw angular velocity, is the velocity of the trajectory reference point, v y is the vehicle lateral velocity, v x is the vehicle longitudinal velocity, e d is the lateral error, is the first-order derivative of the lateral error, is the heading error, is the first-order derivative of heading error; Then the error linear differential equation in the natural coordinate system is obtained as follows: ； in, is the system error derivative matrix, A is the error weight matrix, B is the control weight matrix, is the system error matrix, is the control quantity matrix; ； ； ； ； in, is the second-order derivative of the lateral error, is the second-order derivative of heading error, and The corrected cornering stiffness of the front and rear wheels, is the distance from the vehicle's center of mass to the front axle, is the distance from the vehicle's center of mass to the rear axle; Then the discretized error equation is obtained by Euler method, which is expressed as: ； in, For the k +1The system error matrix at discrete time, is the discretized system error weight matrix, is the discretized control weight matrix, For the k The system error matrix at discrete time, For the k Control quantity at discrete moments; The expression of the discrete system is further obtained as: ； ； in, For the k +1 state quantity at discrete moment, For the k The state quantity at discrete moments, For the k The output at discrete moments, is the discretized output weight matrix, is the discretized state weight matrix.

5. The electric vehicle trajectory tracking control method according to claim 4, characterized in that: Step S4 specifically includes: According to the discrete system obtained in step S3, the state quantity, output quantity and control quantity in the prediction time domain are obtained, and the expression is: ； ； ； in, N p For the prediction time domain, N c To control the time domain, is the state matrix in the prediction time domain, , , Respectively represent k +1 discrete moment, k +2 discrete moments, k + N p The state quantity at discrete moments, is the output matrix in the prediction time domain, , , Respectively represent k +1 discrete moment, k +2 discrete moments, k + N p The output at discrete moments, is the control quantity matrix in the prediction time domain, , , Respectively represent k +1 discrete moment, k +2 discrete moments, k + N c -1 control amount at discrete time, T represents transpose; The vehicle trajectory tracking control model is constructed based on the state quantity, output quantity and control quantity in the prediction time domain, where the cost function J The expression is: ； in, Q and R are the state quantity weight coefficient matrix and the control quantity weight matrix respectively, is the relaxation factor, is the relaxation factor weight coefficient; The optimal control problem is transformed into a standard quadratic programming problem, and the vehicle trajectory tracking control model is solved using a quadratic programming solver to obtain the front wheel steering angle and state quantity output by the vehicle trajectory tracking control model; The obtained state quantity is used as the evaluation index, and the speed pause particle swarm algorithm is introduced to the state quantity weight coefficient matrix Q Perform real-time optimization, including: Define the evaluation index function of the speed pause particle swarm algorithm for: ； in, and is the weight coefficient of the evaluation index function; Define the objective function of the velocity pause particle swarm algorithm for: ； in, yes Q The norm of is the regularization parameter, For the k Evaluation index function at discrete moments; For each initialized particle, the vehicle trajectory tracking control model is used to simulate and obtain the objective function value, which is fed back to the particle swarm. The particle speed and position are updated according to the objective function value, and the speed and position are adjusted in real time. Q , and introduces a speed pause mechanism to pause the particle's speed update under preset conditions.

6. The electric vehicle trajectory tracking control method according to claim 5, characterized in that: In step S5, the expression of the feedforward control model is: ； in, is the feedforward front wheel angle, R w is the tire radius, is the vehicle wheelbase, K is the state feedback matrix of the vehicle trajectory tracking control model; Final front wheel turning angle The expression is: ； in, The front wheel steering angle output by the vehicle trajectory tracking control model.

Images