Research On Automatic Path-finding And Dynamic Obstacle Avoidance Of Driverless Vehicles

As the number of cars on urban roads increasing year by year,the situation of road traffic safety becomes increasingly severe.At the same time,traffic accidents happened frequently.Driving safety also needs more and more attention.At the same time,the development of intelligent transportation and the transformation of the automotive industry have provided the production of driverless cars and new ideas for the resolution of these problems.The safety of the obstacle avoidance path determines whether the car can smoothly avoid obstacles on the road.The presence of driverlessness is a great significance to improving road safety.This paper describes and analyzes the research on automatic path-finding and obstacle avoidance at home and abroad.The study of path planning is divided into two kinds,global path planning and local path planning.This article uses the A* algorithm for global path planning and artificial potential field method for local path planning.At first,this paper describes the principle of A* algorithm in detail and establishes a grid environment.According to the result of simulating with traditional A* algorithm,the premature deletion of the potential optimal raster and the scope of the search limited to 8 neighborhoods commonly caused defects in the planned route.Therefore,the number of searchable neighborhoods of the traditional A* algorithm is expanded from 8 to 16.So the search direction becomes a relatively continuous directions.And proposed a bidirectional algorithm to shorten the search time of A* algorithm.Analyzed by simulation results,the final proposed improving A* algorithm optimizes the above shortcomings and achieves optimality in planning time efficiency and route.The global A* algorithm can only plan the optimal route if the global map is known.But it does not apply to temporary obstacles and moving obstacles that are not shown on the map.Therefore,the traditional artificial potential field method is considered to solve the problem of local path obstacle avoidance.Due to the shortcomings of the traditional artificial potential field method such as local optimum,this article additionally considers the relative distance between the driverless car and the predetermined position.When the car reaches a predetermined position,gravity and repulsion are zero at the same time,and the repulsive potential field function of the obstacle is re-improved.At the same time,the new car has fallen into a local minimum range,so as to solve the problem of local optimality and unreachability.In order to more accurately describe the real scene of the car driving on the road using the potential field method,add a velocity repulsive force field in the same direction as the obstacle repulsion to the moving obstacle,and make the article model closer to the real state.Through simulation,the improved artificial potential field method is verified,the grid environment is improved,and the artificial potential field of A* algorithm is dynamically weighted.After the simulating and comparing,results show that the scope of the algorithm search is smaller and the planning route is more secure and the distance is increased.This paper integrates improved artificial potential field method and A* algorithm and uses ROS platform to simulate the final path planning algorithm.The experiment confirmed that the path planning algorithm can smoothly avoid temporary obstacles and dynamic obstacles not present in the map.