Steady-State Signal Control for Urban Traffic Networks

A typical reason of local traffic congestion is that the distribution of traffic flows in the network is unbalanced, i.e., traffic congestion occurs in some links, but there is still enough space not being sufficiently utilized in other links. Therefore, the consensus notion, extensively investigated in multi-agent networks, is introduced in traffic control, and the steady-state signal control approach is presented, the idea behind which is that traffic states in network links asymptotically converge to given steady-state values, implying balanced distribution of traffic flows is realized. A signal control model of the network is first proposed, which is a linear time-varying system with discharging proportions of internal network links as control variables. Furthermore, analytic relation between steady-state values and inputs of the network is established based on matrix theory, based on which the steady-state signal control law is provided, balancing time-varying traffic flows in the network. At last, simulation investigations are conducted in VISSIM software, compared with fixed-time signal control approach in a real-world topological network in Beijing of China, which shows that the proposed signal control approach can improve performance indices of the network and then reduce local traffic congestion.