基于灰色定性理论的无陷阱人工势场  

Artificial potential field without traps based on grey qualitative theory

作　　者：李书杰[1] 王鹏[1] 陈宗海[1] 

机构地区：[1]中国科学技术大学自动化系,合肥230027

出　　处：《控制与决策》2013年第6期879-883,共5页Control and Decision

基　　金：国家自然科学基金项目(61075073)

摘　　要：针对传统人工势场中存在局部陷阱问题,提出一种基于灰色定性理论的人工势场算法.首先将环境中自由空间分解为一组凸多边形,以凸多边形的顶点和邻接关系作为关键信息,并分别构成灰色定性基本元和灰色定性关系,由灰色定性关系推理从起始点到目标点需经过的凸多边形序列,再用广义白化函数计算凸多边形序列中的势场.理论分析和实验均表明该算法能够确保机器人在有限的时间内安全到达目标点.@@@@Aiming at the local traps problem in the traditional artificial potential field, an artificial potential field algorithm based on grey qualitative theory is proposed. The free space of environment is decomposed into series of convex polygons by using the convex decomposition algorithm. Vertices of convex polygons and the adjacency relationship between them are taken as the key information of environment, which are defined as the grey qualitative fundamental element and the grey qualitative relationship respectively. The sequence of convex polygons from start point to target point is deduced by grey qualitative relationship, and then the potential field in those convex polygons is calculated by generalized whitening function. Theoretical analysis and the experiment verify that the proposed algorithm can ensure the robot reach the target safely in finite time.

关 键 词：灰色定性 人工势场 路径规划 移动机器人 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]