Event control for deadlock avoidance in automated guided vehicle systems

This paper deals with deadlock problems of zone-control automated guided vehicle systems (AGVSs). We assume that a discrete event dynamical model, whose state provides information on the current interaction vehicles-zones, describes the system dynamics. Using digraph tools to characterize and to detect circular wait conditions, a policy is introduced to inhibit and to enable some properly identified events. Such a policy realizes a closed loop control that is effective also in bidirectional guidepath systems where deadlocks and collisions often occur. A case study shows the effectiveness of the proposed approach.