Research On Local Obstacle-avoidance Of Autonomous Vehicle In Front Vehicle Entry Scenes

China’s car manufacturing industry has boomed in recent years,and the country’s car ownership is expected to surpass that of the United States by the end of 2021 to become the world’s largest country with car ownership.However,road congestion and traffic safety accidents caused by more and more vehicles on the road are also increasing year by year.Therefore,the autonomous driving technology of vehicles has gradually become a research hotspot in the current automotive field.Local path planning,as an important research content in the path planning of obstacle avoidance for autonomous driving,plays a decisive role in the realization of obstacle avoidance for autonomous driving vehicles.Advanced local path planning requires not only the hardware support of on-board sensors,but also a high-performance path planning algorithm.The sensor system collects the road information and obstacle information of the surrounding environment in real time and passes it to the decision layer,and obtains a path curve with high fluency and practicability through the path planning algorithm,so as to achieve the purpose of local obstacle avoidance.In this paper,the classification of specific traffic conditions and their components of domestic road traffic dangerous scenes are studied,and on this basis,a new local obstacle avoidance path planning method based on a new improved artificial potential field algorithm is proposed.This method before the car into a dangerous traffic driving behavior as the research scene,in constructing the safety in the process of a lane change lane changing model and dynamic security constraint model this scenario constraint conditions,focuses on analyzing the logic of the driver’s habit and decision-making features,limber cut scene according to the temporal logic sequence can be divided into "start-lane changing-end" phase three child scene,The car body seven-tuple parameter model in static environment is established.On the basis of the analysis of lane change dynamics conditions,the trajectory of obstacle avoidance was simulated and planned based on the quintic polynomial,and the vehicle mass point model was established.Then,the speed impact factor was introduced to evaluate the safety of lane change path,and the vehicle danger level model was constructed under the speed impact factor index.Then,an optimal evaluation model of the local obstacle avoidance curve of automatic driving was proposed based on three indexes,namely curvature,centripetal acceleration of vehicle body,and algorithm operating efficiency.Based on the analysis of the local minimum in the path planning of the classical artificial potential field method,a new path planning method with improved artificial potential field was proposed,and the effectiveness of the improved artificial potential field algorithm was verified and its dynamic performance was optimized and simulated.In the simulation experiment,Car Sim software was used to realize the real-time monitoring of vehicle body kinematics state in the process of obstacle avoidance,and curve description of various dynamic performance indexes was carried out.The effectiveness of the new improved artificial potential field method in local path planning is verified by comparing and analyzing the evaluation model of obstacle avoidance curve optimization.