不确定环境信息下基于方位关系的路径规划算法  

Path planning algorithm based on directional relationship with uncertain environment information

作　　者：吕妍[1] 陈宗海[1] 

机构地区：[1]中国科学技术大学自动化系,安徽合肥230027

出　　处：《中国科学技术大学学报》2013年第10期782-789,829,共9页JUSTC

基　　金：国家自然科学基金(61075073)资助

摘　　要：提出一种不确定环境信息下的移动机器人路径规划方法.首先对环境特征的位置进行高斯分布估计,获得具有概率分布的特征地图,并对其进行凸多边形剖分;然后建立剖分块方向关系模型,根据剖分块之间的相对方向关系构造剖分块的方向优先级,并应用到Dijkstra图搜索算法中,形成基于方向关系的路径规划算法.仿真实验表明,对于具有不确定信息的特征地图,该算法能够有效生成安全路径,并且与A*算法相比较,有效地降低了不确定信息对路径生成的影响.A method for path planning of mobile robots under uncertain environment information was presented. First, the positions of environment features were estimated with Gaussian distribution, to generate a feature map with probability distribution, and the map was subdivided into convex polygon subdivisions. Then, a model of directional relationship of subdivisions was built, and, according to the relative directional relationship of the subdivisions, the priorities of subdivisions were constructed and applied into Dijstra graph search algorithms, and thus a path planning algorithm based on directional relationship was built. Simulation experiments show that the algorithm can generate a safe path by using a feature map with uncertain information, and that in comparison with A ~ algorithm, the algorithm effectively reduces the impact of uncertain information on path generation.

关 键 词：移动机器人 不确定环境信息 路径规划 凸多边形剖分 方向关系 DIJKSTRA算法 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]