Research On Safety Navigation Method Of Autonomous Vehicle In Complex Environment

At present,the research of autonomous vehicle safety navigation methods mainly focuses on positioning,planning and tracking.Although many autonomous vehicles are equipped with various sensors and algorithms,even have completed a certain kilometer of road survey,the reliability of their safety navigation methods is still an important direction in the future field of unmanned driving,especially running in urban or rural areas or even more complex roads requires higher security and robustness of navigation methods.In the research of autonomous vehicle safety navigation method,because of the openness and uncertainty of road traffic environment,"real-time vehicle obstacle avoidance" has become a key scientific problem.For this reason,based on the National Natural Science Foundation Item‘Online Verification of Vehicle Autonomous Decision-making Safety Based on Formalization and Ad-Hoc Method',this paper carries out the following researches:(1)Aiming at the problem that the reliability of autonomous vehicle safety navigation method can not be guaranteed in practical application and the problem of real-time obstacle avoidance in complex environment,a system framework combining positioning,path planning and tracking is proposed.The framework of the system includes vision system,planning and adjustment system and tracking system.In the vision system,differential GPS method is used to locate the vehicle.The data of millimeter wave radar and mobileye sensor are processed by Dspace controller and Simulink module to obtain environmental information.In the planning adjustment layer,the feasibility of the path is judged by the improved artificial potential field method,and this method is used to adjust the path in real time.In the tracking layer,the driver preview model is used to track the planned path.In order to achieve a more stable control effect of the whole system,a double closed-loop PID controller with path closed-loop and lateral preview deviation closed-loop is proposed.The results of simulation and real vehicle validation show that autonomous vehicle can locate autonomously,track and improve the ability of real-time obstacle avoidance,and enhance the reliability of autonomous vehicle safety navigation method under the framework of the system.(2)On the basis of the system framework,in order to improve the speed of re-planning at the planning level,the bidirectional RRT algorithm is combined with the artificial potential field method,and the simulation results are compared in different environments.Through the analysis of the calculation results,it can be seen that the improved algorithm makes the planned path smoother and the planning speed is faster.