Research On Path Planning And Obstacle Avoidance Technology Of Autonomous Vehicle

Path planning is a focus and hot spot in the field of intelligent robot, and it is also an important aspect to realize artificial intelligence. According to the needs of the subject, autonomous vehicle is treated as research object. Path planning algorithm of autonomous vehicle is studied, especially on artificial potential field algorithm and RRT algorithm. Aiming at the limitation and existing problems of the two algorithms, the corresponding improvement methods are put forward. At the same time, real-time obstacle avoidance problem in the dynamic environment is studied using artificial potential field method for path planning.The main parts of the paper include the following sections:Firstly, the path planning algorithm-artificial potential field method is research focus. Introduce the classical artificial potential field method, and analyze its advantages and disadvantages. On the basis of theoretical analysis, the resetting of the potential field function improvement and setting temporary sub-targets are to solve goals unreachable problem and the local minimum problem.Secondly, the RRT algorithm is studied for solving integrity constraints of path planning problem. Describe the principle of RRT algorithm, and analyze its inherent defects. According to the randomness and planning path is not an optimal problem, rolling windows and gravity function of potential field are proposed. The methods implement the RRT algorithm application in unknown environment and planning path is close to the optimal. The simulation proves the reliability and real-time performance of methods.Finally, real-time obstacle avoidance of artificial potential field method in path planning problem is studied in dynamic environment. To solve in local minimum problem of artificial potential field method in path planning, virtual autonomous vehicle by projection method building is put forward. Autonomous vehicle follows the obstacles to avoid obstacle in effect of virtual autonomous vehicle. After escaping the trap, Autonomous vehicle switches back to potential field method. Combining the behavior towards the goal and walking along the wall behavior, a real-time obstacle avoidance algorithm based on behavior is put forward. In order to guarantee the effectiveness of the algorithm in dynamic environment, real-time obstacle avoidance algorithm is improved. Using repulsion potential field of dynamic obstacles, autonomous vehicle movement direction is timely adjusted, to keep the safe distance and obstacles. Simulations verify the effectiveness of the algorithm.