期刊名称：International Journal of Advanced Robotic Systems
印刷版ISSN：1729-8806
电子版ISSN：1729-8814
出版年度：2005
卷号：2
期号：3
DOI：10.5772/5787
语种：English
出版社：SAGE Publications
摘要：This paper examines path planning strategies in partially unknown dynamic environemnts and introduces an approach to learning innovative routes. The approach is verified against shortest path planning with a distance transform algorithm, local and global replanning and suboptimal route following in unknown, partially unknown, static and dynamic environments. We show that the learned routes are more reliable and when traversed repeatedly the robot's behaviour becomes more predictable. The test results also suggest that the robot's behaviour depends on knowledge about the environemnt but not about the path planning strategy used.
关键词：path planning; mobile robots; dynamic environment; robot learning
Loading...
