Lianhao Zhang∗,∗∗ Zhenyu Liu,∗ and Xiaohong Qin∗∗
Characteristic statistics, obstacle avoidance, path planning, mini-mum bounding rectangle, MBR-PRM
Path planning with obstacle scenes is a rapidly growing research area about robots, and most algorithms focus on their own improvement and optimization. However, there is little research on obstacles, and when the obstacles in the scene are changed, some parameters in the existing algorithm have to be set manually, so it is difficult to adapt to the transformation of complex obstacle scenarios, and the chance of success is slim. This paper is concerned with the statistic analysis of the number, shape, complexity, distribution and other characteristics of obstacles in the work scene. On that basis, it forms a corresponder function between the characteristics of obstacles and the main parameter of the improved probabilistic roadmap algorithm, adjusting the sampling area and numerical size of sampling point without manual setting. In addition, considering the time consumption, path complexity and other factors, which combines the minimum bounding rectangle algorithm in the obstacle area, the relatively short path is selected and the success rate of path planning is guaranteed. Compared with other commonly used algorithms, the novel algorithm proposed in this paper is more general and adaptive to the changeable obstacle scene, and the experimental results also demonstrate that it can effectively avoid obstacles and significantly improve the evaluation criteria of path planning.
Important Links:
Go Back
