Using Anytime Planning for Centralized Coordination of Multiple Robots in Real-Time Dynamic Environments

Gabriel J. Ferrer, Glenn S. Wasson, James P. Gunderson, and Worthy N. Martin, University of Virginia

We are investigating the use of planning in multi-robot, real-time, dynamic environments. Our paper discusses an anytime planning algorithm we have devised for the particular problem of replanning. Our anytime algorithm begins by repairing a restricted part of the plan without altering the rest of the plan. The part of the plan that can be replanned is then progressively increased until the algorithm is interrupted or the entire plan has been replanned.