Traffic-responsive signalling control through a modular/switching model represented via DTPN

Urban signalized traffic areas are considered in this paper, with the aim of minimizing congestion situations via a traffic-responsive signalling control procedure founded on a hierarchical Petri net (PN) representation of the system. The higher level of the PN representation consists of net modules, each one representing an intersection, a road, a signal staging, etc.; the description of each module in terms of deterministic-timed Petri nets (DTPN) is given at the lower level. Such a representation leads to a corresponding two-level control procedure. The high-level control system, which acts over the modular representation, switches among internal module structures so as to modify some parts of the model of the traffic system (e.g., signal plans, turning rates, etc.), depending on both state and time. The low-level control system, which acts over the DTPN representation, optimizes the performances of the traffic system, by solving a mathematical programming problem which minimizes the number of vehicles in the system. In the paper, the adopted model of the signalized urban area is briefly presented, and the two-level representation and the control system are described in details.