Dynamics of Cyber-Physical Systems for Intelligent Transportations

Cyber-Physical System is a complex system that integrates perception, computing, communication and control to construct the mapping, interaction and coordination of human, machine, object, environment and information in cyber and physical space. In transportation cyber-physical system, the topology of road network is extremely complex, vehicles are widely distributed, and information changes rapidly. With the development of embedded systems and communication technology, interactions between vehicles and other devices are becoming faster and more reliable, making dynamic route guidance possible. This paper proposes a dynamic route guidance method named Q-learning Rerouting (QRR), which updates Q-value table by periodically acquiring real-time average travel time as penalty of actions and guides vehicles according to maximized value function strategy. Two common scenarios in intelligent transportation are built: congestion caused by vehicle failure and speed limit on waterlogged roads. Compared with static route guidance, QRR makes full use of road capacity and allows vehicles move faster.