Hybrid Path Planning Based On Particle Swarm Optimization And Improved Path Artificial Potential Field For Mobile Robot

The mobile robot path planning is a key research area in robotics. Based on the realistic demand, A hybrid path planning approach which integrates Particle Swarm Optimization(PSO) and improved Artificial Potential Field(APF) is proposed in partly unknown and dynamic environment. At first, a global path is obtained by improved PSO method; and then a local obstacle avoidance algorithm based on artificial path potential field associated with the global path is presented, which approach the nearly optimal path.Main contents of this thesis are as follows:1. As the global planner of the hybrid approach, this paper uses the PSO in the improved Grid Method. Firstly, it get the effective vertexes of grid-obstacle which the obstacle in the environment is represented by Grid Method, to generate encoding sequence. After that, it applies the PSO to the improved grid encoding sequence to get the best robot global path. It can easily overcome the traps of obstacles in path planning, and makes the path planning more simple and effective, it also makes the speed of PSO convergence faster. The random factor in the algorithm can avoid it traps into local optimum.2. As the local planner of the hybrid approach, this paper proposed an improved APF algorithm. Except the traditional local potential field, the improved algorithm also considers the potential field of velocity and acceleration. With those factors, it can be fluently applied to collision-free motion planning in dynamic environment. Also, the improved APF algorithm has a consideration of the attraction of an expected path, we call it’line potential field’. Because of the existence of line potential field, the robot fully uses the already known information. This algorithm works well with the global planner in the hybrid approach.3. Combined the idea of global planner and local planner, this paper gives the process of the hybrid approach. We apply it to the AS-R mobile robot in our lab, the experiment convert the virtual force of the APF to the control of our robot’s acceleration. To solve the problem of unreachable acceleration’s control point in crawler robot’s control, an expectation acceleration method is proposed, it uses the suboptimal control point to substitute the unreachable part. The simulation results and physical experiment results proved our approach is simple and feasible and it is applicable to the actual environment.