Research On Global Dynamic Path Planning Based On Improved A* Algorithm

The application of intelligent car in life is becoming more and more common.Due to the complexity of road environment and the variability of obstacles,whether the intelligent car can reach the designated position quickly and safely has become one of the important standards to measure the level of intelligent car.In the process of path planning,the intelligent car mainly carries out path search based on the path planning algorithm.Currently,many path planning algorithms are relatively mature and have their own advantages,but they all have certain defects.This paper mainly studies A* algorithm,optimizes its defective parts,improves its planning ability,and carries out simulation and experiments to prove its feasibility.The research of this paper is mainly as follows:(1)First of all,aiming at the defect of A* algorithm with large amount of computation in path planning,the heuristic function of A* algorithm is improved,and the path search strategy of A* algorithm is proposed by using jump point method: A pre-processing process is added before A* algorithm path planning,and A* algorithm is combined with skip method to access the contents of multiple nodes on the map with fewer times of A* iteration,which reduces A lot of unnecessary calculations and thus shortens the time of path search.The comparative simulation results show that the calculation amount of the improved A* algorithm with jump method is greatly reduced and the planning efficiency is improved.(2)because of the influence of planning path by A * algorithm is A turning point more cars driving efficiency,aiming at the problem of A * algorithm path does not smooth interpolation point method is put forward for its optimization: use dynamic programming ideas for A given weighted graph of the shortest path between more than A turning point in,remove excess point and inflection point of the path,merge collinear node,the actual path more in line with the intelligent vehicle trajectory.Through simulation comparison and verification,after the interpolation method is optimized,the redundant nodes in the path are eliminated,and the motion trajectory of the intelligent car is smoother.(3)A* algorithm has weak planning ability in the face of dynamic obstacles.After improving its computing efficiency and path smoothness,this paper combines the optimized algorithm with dynamic window method to design A global dynamic path planning algorithm so that the intelligent car can realize real-time obstacle avoidance in the process of driving.The simulation results show that the efficiency of the proposed algorithm is greatly improved,the path is smooth and effective,and the algorithm has the ability of dynamic obstacle avoidance.(4)At the end of this paper,the improved algorithm is verified by experiments: in the unknown obstacle environment,the sensor carried by the intelligent vehicle is used to transmit the surrounding environment information to the ROS system,and the three-dimensional visualization software RVIZ is used to establish the spatial model.According to the environment map model,the intelligent car can avoid obstacles and reach the designated position smoothly by the global dynamic path planning algorithm.The success of road environment experiment of intelligent vehicle effectively verifies the theoretical reliability.