Petri net approach for deadlock prevention in robot planning

This paper provides a strategy for supervising the motion of some mobile robots that evolve in the same environment. Some regions of the environment are assumed to have a limited capacity in terms of the number of robots that can simultaneously occupy them, and a set of possible trajectories is available for each robot. The solution comprises the construction of a specific Petri net model for the available trajectories, and the use of resource-allocation techniques based on restricted-capacity regions and on deadlock-free execution.