Research On Trajectory Tracking Of Autonomous Driving Vehicles Based On Model Predictive Contro

It is particularly important that the unmanned vehicle tracks the pre-planned trajectory accurately and smoothly in the process of automatic travel.In the process of tracking,obstacles sometimes appear at local locations,and they pose a great threat to the safety of the unmanned vehicle.Therefore,it is necessary to ensure that the unmanned vehicle travels strictly according to the reference trajectory and that the unmanned vehicle can adopt an emergency collision avoidance strategy to effectively avoid the obstacles when it encounters them.In this paper,the trajectory tracking and local collision avoidance of self-driving vehicles are investigated based on model predictive control algorithms with an unmanned vehicle as the object of research,and the main work is as follows:Aiming at the problem of tracking accuracy and stability degradation when the unmanned vehicle tracks a complex,non-single reference trajectory and there are no obstacles on the trajectory,this paper proposes a method that combines tyre model and unmanned vehicle dynamics,establishes a three-degree-of-freedom dynamics model of the unmanned vehicle that considers tyre lateral deflection angle,and adds dynamics constraints to this model,which fully considers the tyre forces in the unmanned vehicle driving The tyre lateral force caused by the tyre is fully considered in the model,and a model predictive trajectory tracking controller is designed based on the model.The results show that the designed trajectory tracking controller can meet the requirements of trajectory tracking.A local collision avoidance strategy is proposed for the problem that obstacles can prevent the unmanned vehicle from continuing tracking when there are obstacles on the tracking reference trajectory.On the one hand,for static obstacles,a local obstacle avoidance reprogramming algorithm is proposed,and the obstacle size coordinates and obstacle avoidance function functions are improved therein,by doing expansion and segmentation of the obstacle size,establishing a mathematical model of the obstacle avoidance weight factor based on the least squares method,and modifying this obstacle avoidance weight factor in real time according to the given vehicle speed and road adhesion coefficient.On the other hand,for dynamic obstacles,a braking and following algorithm based on model prediction control is proposed.The algorithm first obtains the speed and position information of the vehicle in front through Kalman filtering,and solves the speed and position of the vehicle in real time,and predicts the relative motion state of the two vehicles in the future period through the model prediction control algorithm,so that the unmanned vehicle can do following while keeping a certain distance from the vehicle in front The vehicle can follow the car while maintaining a certain distance from the car in front.Finally,the experimental verification is completed.Firstly,the trajectory tracking experiment of the unmanned vehicle was carried out in the absence of obstacles to verify the effectiveness of the model prediction trajectory tracking controller designed in this paper.Secondly,the local trajectory replanning algorithm based on the improved obstacle avoidance function and the braking following algorithm based on the Kalman filter algorithm designed in this paper are experimented in real vehicles in the presence of obstacles,and finally the effectiveness of the local collision avoidance strategy proposed in this paper is also verified.