In this paper weshowthat parallel search techniquesderived fromtheir sequential counterparts can enable the solution of instances of the robot motionplanningproblemthat are computationallyinfeasible on sequential machines.Wepresent a simpleparallel version of a robot motionplanning algorithmbasedon "quasi best first" search with randomizedescape from local minimaand randombacktracking and discuss its performanceon two problem instances, one of which was computationaUyinfeasible on a single processor of an nCUBE21 multicomputer. Wethen discuss the limitations of parallel robot motion planning systems, and suggest a course for future work.
Daniel J. Challou, Maria L. Gini, Vipin Kumar