Online Petri net based algorithm for planning and controlling mobile robots

The paper presents a procedure for planning and controlling a team of identical mobile robots such that a set of target regions are reached. We consider a partitioned environment cluttered with a set of obstacles that randomly change their positions. Our approach abstracts both the control capabilities of robots and the information on obstacle locations into probabilistic discrete Petri net models, and it uses an online algorithm for planning and adjusting the sequences of regions the robots should traverse. At each iteration of the online algorithm we solve a linear programming problem that optimizes the robot paths by weighting the probabilities of following a sequence of partition regions and the steady-state probabilities of encountering obstacles along the resulted sequences. The approach is implemented as a fully automated Matlab package.