Home /Research /ARTIC ULAT ED ROBOTS MOTION PLANNI NG USING FO RAGING ANT STRATEGY
OTHER

ARTIC ULAT ED ROBOTS MOTION PLANNI NG USING FO RAGING ANT STRATEGY

Mohd Murtadha Mohamad

Year
2008
Citations
2

Abstract

A bst ract : Ma ny different approaches to tac kle the pro blem of motion pla nn ing fo r articu late d robots in an env ironment w ith obstacles based on random sampling have been proposed. One popular ap proac h is called single-query bi-directional motion pla nning w ith a lazy co llision checking pro ba bilistic ro adm ap (S BL-P R. 1.) . However , the performa nce of th is met hod is sub -op tima l in terms o f the num ber of configurations ge nerated, length of path, amount of col lisio n chec king and computatio nal time . To improve the performance, those as pec ts must be co ns idered furt her as they are inter-related with eac h other. A novel mo dification o f SBL­ PRM that dec reas es the size of excessive configurations in th e ro ad map, by incrementa lly bu ildi ng a one-t ree st ructu re orig inat ing from the sta rt configu rat ion, is prese nted. This approach, the single-q uery u nidi rectio na l ap proac h with lazy co llision c hec king (S UL-P R.\1), has ex perimental ly sho w n to be equal to the SBL· PR.\'!. How ever, there still e x ists ge nerated configurations th at were exc luded from the success ful path . The generat ion o f these unco nsume d co nfigurations co rres pond ing to the tree structure has poin tless ly utilized th e com putationa l resource s and affected the planni ng time. Hence, a new method o f configurat ion ge nerat ion a long w ith a novel searc hing style is devised. An alte rnat ive sea rch approac h usi ng ant behav iour in a robotics app lication is a pplied. This paper proposes a nove l searc h technique, the F- Ant a lgorithm, in orde r to find a re liabl e path betw een the initi al con figuration and the goal con figuration of the articulated ro bot. This no ve l algorithm, taking two in put con figu ratio ns, e xplo res th e robot's free s pace by build ing up a u nid irect ional searc h be ginn ing at the init ial configuration. The planner sam ples the free co nfigurat ion repeti t ively in the ne igh bourhood within the rad ius of the c urre nt co nfigurat ion, a nd tests th e edge for a collisio n-free pat h between the new sampled configuratio ns, until it is con nected to th e goal configuration. Simu latio n and experimenta l co mpariso ns o f F-Ant and SBL- PR M have bee n con d ucted, sho w ing t he per forma nce differences betw een these two tec hniques.

Keywords

Path (computing)AlgorithmPhysicsCombinatoricsMathematicsGeometryComputer scienceComputer network

Related papers

Browse all OTHER papers