Previous [ 1] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] [ 8] [ 9] [ 10] [ 11] [ 12] [ 13] [ 14] [ 15] [ 16] [ 17] [ 18] [ 19] [ 20]

@

Journal of Information Science and Engineering, Vol. 23 No. 2, pp. 525-538 (March 2007)

An Incremental Learning Approach to Motion Planning with Roadmap Management*

Tsai-Yen Li and Yang-Chuan Shie
Department of Computer Science
National Chengchi University
Taipei, 116 Taiwan

Traditional approaches to the motion-planning problem can be classified into solutions for single-query and multiple-query problems with the tradeoffs on run-time computation cost and adaptability to environment changes. In this paper, we propose a novel approach to the problem that can learn incrementally on every planning query and effectively manage the learned road-map as the process goes on. This planner is based on previous work on probabilistic roadmaps and uses a data structure called Reconfigurable Random Forest (RRF), which extends the Rapidly-exploring Random Tree (RRT) structure proposed in the literature. The planner can account for environmental changes while keeping the size of the roadmap small. The planner removes invalid nodes in the roadmap as the obstacle configurations change. It also uses a tree-pruning algorithm to trim RRF into a more concise representation. Our experiments show that the resulting roadmap has good coverage of freespace as the original one. We have also successful incorporated the planner into the application of intelligent navigation control.

Keywords: incremental learning, motion planning, probabilistic roadmap management, reconfigurable random forest, planning for dynamic environments

Full Text () Retrieve PDF document (200703_11.pdf)

Received March 31, 2005; revised November 11, 2005; accepted December 22, 2005.
Communicated by Chin-Teng Lin.
* This research was funded in part by the National Science Council of Taiwan, R.O.C., under contracts No. NSC 93-2213-E-004-001 and NSC 94-2213-E-004-006.