摘要
研究了环境未知情况下的移动机器人路径规划问题,将快速扩展随机树(RRT)算法与基于滚动窗口的路径规划相结合,提出一种新的移动机器人路径规划算法,克服了RRT算法通常只能在已知环境中进行移动机器人路径规划的限制。规划时只考虑窗口环境地图,提高了RRT算法规划效率,保证了算法的实时性。针对RRT算法路径规划缺乏确定性的问题,结合人类经验及模糊控制理论,以概率来随机取点,并引入启发式估价函数,使随机树易于朝目标点方向生长。同时运用回归分析生成新节点,增强了算法搜索未知空间的能力,避免了可能产生的局部极小。最后仿真实验验证了该算法的有效性。
The mobile robot path planning in an unknown environment is studied.The rapidly-exploring random tree（RRT） algorithm is combined with the rolling path planning in the planning,and a novel path planning is proposed.It is performed on-line in a rolling style,so that the RRT algorithm can be used in path planning not only when the environment is known but also unknown.Only the local environment map is calculated in planning so as to improve the planning efficiency,and the real time is guaranteed.Owing to lack of certainty in path planning of RRT,the random configuration using both fuzzy control theory and human experience is forced to coincide with the desired area with some finite probability.The algorithm has introduced the heuristic evaluation function so that the exploring random tree will grow in the direction of target point.The regression analysis,which avoids local minima,enhances the capability of searching an unknown space.The method effectiveness is verified by simulation results.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2010年第5期642-648,共7页
Journal of Nanjing University of Science and Technology
基金
国家"863"计划资助项目(2006AA04Z238)
南京理工大学国防科技工业研究生创新基金(2007060007)
