Home /Research /Multi-robot planning under uncertainty with communication: a case study
SWARM

Multi-robot planning under uncertainty with communication: a case study

João Messias, Matthijs T. J. Spaan, Pedro U. Lima

Year
2010
Citations
4

Abstract

Although Dec-POMDP techniques can be useful to mod-eling a wide range of problems, their practical application is limited by the inherent computational complexity of the algorithms currently available to solve such models. The ap-plication of these techniques is typically restricted to theo-retical examples. This work studies the application of a par-ticular type of Dec-POMDP (a multiagent POMDP) model to solve a simple task in a realistically simulated robotic soccer environment. The multiagent POMDP exploits the availability of a communication channel, as is often the case in multi-robot systems. The necessary constraints on the problem are identified, and the steps taken to accomplish efficient cooperative behavior within real robot middleware are explained. Finally, results are presented for the proposed task that highlight further possible improvements.

Keywords

Partially observable Markov decision processComputer scienceTask (project management)RobotArtificial intelligenceExploitSimple (philosophy)Distributed computingMachine learningEngineering

Related papers

Browse all SWARM papers