Research On Local Trajectory Planning And Tracking Control Algorithm Of Autonomous Vehicles

The trajectory planning and tracking control algorithm is an important part of the autonomous driving system.Trajectory planning is the bridge between the perception module and the tracking control module,and the tracking controller provides instructions for the execution module of the autonomous vehicle.With the diversification of driving scenarios and the increasing requirements for vehicle control accuracy,the demand for the performance of trajectory planning and tracking control algorithms is getting stronger and stronger.In order to improve the performance of autonomous vehicle trajectory planning and tracking control algorithm.This thesis combines the traffic risk field theory with the state grid method,proposes a local trajectory planning strategy in the scenarios of curved and no-signal intersection scenarios,and establishes a trajectory tracking controller based on the model predictive control(MPC)algorithm.The specific contents are as follows:(1)Fully consider the combined role of the driver and the elements of the road environment,the local trajectory planning algorithm of autonomous vehicles is proposed by combining the driving risk field theory with the state grid method.Firstly,under the Frenet coordinate system,the polynomial method is used to generate the vehicle motion trajectory.Secondly,a trajectory feasibility check strategy is formulated to eliminate trajectories that do not meet the constraints of vehicle dynamics.Finally,the trajectory quality evaluation function is constructed with driving risk and comfort as the core to provide data basis for selecting the optimal trajectory.(2)For the curved road and traffic intersection scenarios without signal lights,the local trajectory planning models are established.In the curve environment,the proposed trajectory planning algorithm is applied to this scenario.In the environment of intersection without signal lights,a decision-making algorithm is designed by using the relationship between vehicle driving risk and risk threshold.By obtaining the risk value between the self-driving vehicles and other vehicles,if it is less than the threshold,let the vehicle pass through the intersection.Instead,it will slow down and wait for the next opportunity to pass the intersection.Experimental results prove that the planning algorithm proposed in this thesis can give a reasonable driving scheme.(3)Based on the vehicle dynamics model,the MPC trajectory tracking controller is established.Firstly,according to the vehicle dynamics model,the discretization and linearization methods are used to obtain the prediction model,and the objective function with constraints is designed,so as to complete the design of the trajectory tracking controller.Then,through the simulation and comparison experiments of front-wheel feedback control algorithm,linear quadratic regulator control algorithm and MPC algorithm,it is verified that the MPC algorithm has better control accuracy than the other two control algorithms.Finally,the influence of the prediction time domain and control time domain parameters on the tracking effect of the controller in the predictive controller is analyzed,and a dual-time domain parameter update strategy is proposed.The experimental results show that the control accuracy of the tracking controller is significantly improved after optimization.(4)Car Sim and Matlab/Simulink were used to build a joint simulation platform to verify the tracking effect of the tracking control system designed in this thesis.The vehicle model is built in Car Sim,and the S function in Simulink completes the code programming of each module,and then completes the construction of the joint simulation platform.Under the condition of double line shift,simulation experiments verify that the trajectory tracking controller based on MPC has a good tracking effect.Finally,the planning layer is added on the basis of the original controller to form a double-layer control system with trajectory replanning and tracking control.Experimental results indicate that the control system can stably track the reference trajectory in static and dynamic obstacle environments.