A General Framework for Intersection Traffic Control With Backpressure Routing

Traffic congestion has long been a worldwide difficult problem in metropolitan transportation networks, incurring tremendous time waste and exhaust pollution. Extensive efforts have made to address this problem, among which traffic control at intersections is known be especially crucial while challenging. Also, it helpful the recent advent of autopilot technologies by enhancing schedule accuracy reducing infrastructure cost. In practice, however, existing researches can hardly work well pervasive manner since they are essentially limited two ideal assumptions: 1) each intersection comprises four ways; 2) way homogeneously composed lanes. Through an in-depth examination their basic models, we find that two-fold assumptions largely compelled surprisingly high complexity converting practical topology into theoretical conflict graph. Moreover, works seldom consider modeling complex scene cooperative multiply intersections. Driven above understandings, our first effort towards general framework (for handling multiple-way heterogeneous intersections) carefully transform homomorphic, regular graph suited further processing with affordable complexity. Besides, maximum weight independent set (MWIS) based approach proposed minimize average waiting vehicles isolated intersection. addition, apply backpressure-based algorithm optimize global whole road network. Simulation results demonstrated achieved merely 80 seconds traditional light reaches 370 seconds.