Planning multiple paths with evolutionary speciation

被引:79
作者
Hocaoglu, C [1 ]
Sanderson, AC
机构
[1] SUNY Stony Brook, Dept Elect & Comp Engn, Stony Brook, NY 11794 USA
[2] Rensselaer Polytech Inst, Dept Elect Comp & Syst Engn, Troy, NY 12180 USA
基金
美国国家科学基金会;
关键词
evolutionary speciation; path planning;
D O I
10.1109/4235.930309
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper demonstrates a new approach to multidimensional path planning that is based on multiresolution path representation, where explicit configuration space computation is not required, and incorporates an evolutionary algorithm for solving the multimodal optimization problem, generating multiple alternative paths simultaneously. The multiresolution path representation reduces the expected search length for the path-planning problem and accordingly reduces the overall computational complexity. Resolution-independent constraints due to obstacle proximity and path length are introduced into the evaluation function. The resulting path-planning system has been evaluated on problems of two, three, four, and six degrees of freedom, The resulting paths are practical, consistent, and have acceptable execution times. The system can be applied for planning paths for mobile robots, assembly, and articulated manipulators. Generation of multiple alternative paths is an example of multimodal search and, in our previous work, a new approach to multimodal function optimization has been developed using a genetic algorithm (GA) with minimal representation size cluster analysis. This multiple-population GA identifies different species and is used as the basis far an evolutionary multipath-planning algorithm which generates multiple alternative paths simultaneously. The multipath algorithm is demonstrated on a number of two-dimensional path-planning problems.
引用
收藏
页码:169 / 191
页数:23
相关论文
共 61 条
[51]   A SURVEY OF MOTION PLANNING AND RELATED GEOMETRIC ALGORITHMS [J].
SCHWARTZ, JT ;
SHARIR, M .
ARTIFICIAL INTELLIGENCE, 1988, 37 (1-3) :157-169
[52]  
Schwartz JT., 1983, Adv. Appl. Math., V4, P298
[53]  
SEGEN J, 1981, CMURITR822
[54]   ALGORITHMIC MOTION PLANNING IN ROBOTICS [J].
SHARIR, M .
COMPUTER, 1989, 22 (03) :9-20
[55]  
SHARIR M, 1989, P IEEE S ROB AUT OCT, P9
[56]  
SHIBATA T, 1993, P IEEE 8 INT S INT C, P182
[57]   FORMAL THEORY OF INDUCTIVE INFERENCE .I. [J].
SOLOMONOFF, RJ .
INFORMATION AND CONTROL, 1964, 7 (01) :1-+
[58]   Motion planning for carlike robots using a probabilistic learning approach [J].
Svestka, P ;
Overmars, MH .
INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH, 1997, 16 (02) :119-143
[59]  
WALL M, 1996, GALIB AC PLUSPLUS LI
[60]  
XIAO J, 1997, IEEE T EVOLUTIONARY, V1, P18