Hierarchical and distributed control of networked vehicle systems using discrete event/hybrid system theory

As more and more vehicles being added to the road every day, it is important to investigate new ways to meet the growing demand on ground transportation systems. One way is to adopt a fully Networked Vehicle System (NVS), where the information is exchanged and control is applied to ensure smooth traffic flow. We propose to use discrete event/hybrid system theory to model and control such an NVS in a hierarchical and distributed fashion. Control is achieved at three levels: vehicle level, infrastructure level, and system level, by vehicle supervisors, infrastructure supervisors, and master supervisors, respectively. At the vehicle level, vehicles are modeled as hybrid machines, where both continuous variables and discrete events are considered. At infrastructure and system levels, all vehicles are modeled as finite state machines with parameters, which can be viewed as abstractions of hybrid machines. Control at each level has its own control objectives and is implemented in a distributed fashion.