Research On Local Path Planning And Tracking Control Of Autonomous Vehicles

The development of automobile autonomous driving technology is an effective way to solve a series of traffic problems and environmental problems caused by the increasing number of automobile ownership.It is the development trend of automobile under the development of automobile industry,information technology and artificial intelligence technology.Autonomous driving technology has also gradually become a research hotspot in relevant enterprises,universities and scientific research institutions.Planning and control technology is a key technology in autonomous driving technology,which directly affects the driving safety and comfort of autonomous vehicles.This paper focuses on local path planning and tracking control in autonomous driving technology.In the structured road environment,an optimal obstacle avoidance trajectory is obtained through the local path planning algorithm,and the tracking control algorithm controls the planned trajectory with centimeter-level tracking accuracy,while ensuring the stability of the vehicle driving.The specific research contents are shown as follows:(1)Local path planning.In this paper,we decompose the local path planning problem into two problems: path planning and speed planning.For path planning,its main task is to perform static obstacle avoidance based on the Frenet coordinate system.In this paper,in the path boundary,the path boundary meets the requirements of obstacle avoidance and the vehicle kinematic constraints.For speed planning,its main task is to avoid dynamic obstacles based on the ST diagram.In this paper,we get the rough speed planning results in the ST diagram,and then they satisfy the traffic and vehicle motion constraints based on the nonlinear constraint programming algorithm.Finally,the effectiveness of the path planning and speed planning algorithm in local path planning.(2)Track-tracking control.In order to reduce the complexity of control and improve the timeliness of control,the method of decoupling from longitudinal and transverse control is applied to the tracking control problem of autonomous vehicles.The longitudinal control adopts the position-speed dual PID control algorithm,in which the control amount of the acceleration / brake pedal is determined according to the expected acceleration obtained from the control and the current vehicle speed search calibration table.The lateral control uses an LQR control algorithm based on front-wheel feedback.The LQR algorithm was determined by modeling the vehicle dynamics and eliminating the effects of the pathway via feedforward control.Aiming at the insufficient control accuracy of the LQR controller based on feedforward control,the front wheel feedback control is introduced to improve the control accuracy and ensure the stability of the vehicle.Finally,the simulation algorithm the lateral control and compare the simulation effect of the three controllers.(3)Co-simulation of local path planning and tracking control.The effectiveness of the proposed local path planning algorithm and tracking control algorithm is verified in the Simulink-Carsim Co-simulation platform.The simulation verification results show that the local path planning and tracking control algorithm proposed in this paper can realize the obstacle avoidance and tracking functions in typical scenarios.(4)Finally,the real vehicle algorithm was tested in different typical scenarios,and the test results were analyzed.The test results show that the proposed local path planning and tracking control algorithm can show good results in real vehicles,and the tracking control effect has high accuracy and stability.In this paper,the separate simulation,joint simulation and real vehicle test verification of the proposed local path planning algorithm and tracking control algorithm prove that the proposed algorithm has good obstacle avoidance and tracking function,and can meet the requirements of autonomous vehicles for local path planning and tracking control.