Fuzzy discrete event system modeling and temporal fuzzy reasoning in urban traffic control

In this paper, a DEVS-based fuzzy supervisory controller is applied to an urban traffic control problem also modeled from a discrete event perspective via state flow formalism. Time varying nature of intersections, involvement of human behavior in the system and distributed nature of the problem at large makes the urban traffic control a challenging problem. Specifically, because of this distributed nature, it is important for the decision making paradigm to consider historical values of the linguistic variables as well as their current values. Hence, we establish two novel fuzzy systems which designate cycle time and split time based on temporal fuzzy rules. Temporal quantifiers are proposed, and rules are designed to resolve inconsistency in temporal information. Simulation results present the effectiveness of the proposed method.