Research On Signal Timing Methods For Oversaturated Intersection In Urban Environment

Traffic signal control is one of the most economical and efficient methods to increase the traffic capacity of intersections in urban traffic environment,as well as alleviate traffic congestion and improve vehicles' ride quality.Signal timing can be regarded as a multi-objective optimization problem.In this thesis,the typical cross intersection is taken as the research object.And a multi-goal optimizing function is set up to maximize traffic volume,minimize average waiting time,average number of stops and CO₂'s emission.Under oversaturated traffic conditions,the overflow of traffic stream leads to failure of traditional signal timing methods.In this thesis,signal timing methods used for oversaturated intersection in urban environment are studied.Firstly,by considering the types of vehicles,a signal timing frame is proposed for oversaturated intersection.Traditional signal timing methods are used in the case of spare traffic volume,and oversaturated signal timing methods are used in the case of dense traffic volume.Then one kind of signal timing method is proposed for oversaturated intersection based on NSGA-? genetic algorithm.Genetic algorithm is one of the most efficient means to find Pareto front of multi-objective optimization problems.The aim of multi-objective optimization problem is how to find optimal solutions that evenly distributed in Pareto front as many as possible.And the key is to reduce conflict between population diversity and convergence property.At last,another method is proposed based on dynamic programming theory.The advantage of this method is that it can transformed the procedure of signal timing optimization into several single-step optimal controls.Many performance indicators can be calculated by multiple iterations and decide whether switch or keep states of phases.Besides,the next optimization procedure is adjusted according to the feedbacks.As a result,the traffic signal control system is adaptive in dynamic traffic conditions.Simulations are conducted on the open source vehicular network simulation platoon with real traffic data.The simulation results show that the proposed signal timing methods outperform traditional fixed timing method and Webster-based timing method,in terms of ride quality of intersection.To be specific,on the premise of guarantee traffic volume,they can reduce the average waiting time and number of stops,as well as CO₂'s emission.