Research On Traffic Signal Control Optimization Methods Under Unsaturated State

The rapid development of urban economy and accelerating urbanization process have led to a significant increase in the consumption level of residents,which is directly reflected by the rapid increase of vehicle ownership.Therefore,the traffic load and traffic pressure of urban roads are unprecedented.The contradiction between traffic demand and traffic supply is prominent increasingly,which causes a series of traffic problems,such as traffic congestion,traffic safety,environment pollution and so on.The traffic problems above seriously restrict the development of the city and influence people's travel efficiency.Effective and reasonable signal control scheme of urban intersections can reduce traffic accidents and delay time,increase road capacity and ensure the safe and orderly operation of traffic,which has good social benefits.The rapid development and gradual implementation of ITS put forward higher requirements for the city traffic signal control.Therefore,how to make more rational design and effective optimization of the road traffic signal control schemes through intelligent technology has been attracting the attention of researchers at home and abroad.To optimize the urban traffic signal control scheme and improve the efficiency of road network traffic,based on signal control range,this paper mainly focuses on the urban traffic signal control optimization method under unsaturated state.Advanced intelligent algorithm,combined with time series analysis,clustering analysis,community classification methods theories is employed to conduct the thorough research.The main research contents and results are as follows:(1)Signal control period division based on short-term traffic flow predictionConsidering the advantages of the hybrid prediction method,a short-term traffic flow prediction method based on CEEMDAN-PE-OSELM is proposed.The traffic flow time series are decomposed by CEEMDAN algorithm and several IMF components can be obtained.Then PE algorithm is applied to recombine the IMF components.OSELM prediction models are established for each restructured subsequence by PE algorithm respectively.To obtain the final prediction result,the prediction results by OSELM prediction models are added together.A typical intersection is chosen to verify the proposed hybrid prediction model of which the performance is compared with other similar methods.On the basis of the short-term traffic flow prediction,the Fuzzy C-Means Clustering(FCM)algorithm is improved and a new signal control period division model based on RSAGA-FCM algorithm is proposed,providing the basis for subsequent signal timing scheme optimization.(2)Signal control optimization based on fuzzy compromise programming for isolated signalized intersectionFocused on isolated intersection,traffic capacity,vehicle cycle delay,cycle stops,and exhaust emission are selected as optimization goals to establish a multi-objective function,considering the whole operation efficiency of the intersection comprehensively.For different traffic flow states,fuzzy compromise programming approach is employed to give different weight coefficients to various optimization objectives and convert the multi-objective function to a single objective function.Genetic algorithm is applied to solve the optimization model to obtain the optimized signal cycle and effective green time.Finally,the simulation evaluation of different signal schemes is carried out by VISSIM software.(3)Arterial coordination control optimization considering vehicle arrival and dissipationConsidering the arrival and dissipation characteristics of vehicles between the intersections of the arterial system,two kinds of coordination control optimization models are proposed.Combined with the mixed integer linear programming method,an improved MULTIBAND model considering dynamic queuing dissipation time is put forward.Firstly,the form of the green wave band in MULTIBAND model is optimized.Secondly,the traffic queue characteristics between the upstream and downstream intersections are analyzed,and the dissipation time model of red light queuing is established to improve the MULTIBAND model.Finally,an arterial system is selected for model verification.In addition,a minimum delay arterial coordination control model considering vehicle arrival and dissipation is proposed.Vehicle arrival type is divided into 10 kinds based on the different moment the first vehicle in coordinated phase dissipating from upstream intersection arrives downstream intersection.Several arterial delay models are established for each vehicle arrival type and the function relation between the whole arterial delay and system offsets can be obtained.With the goal of minimizing total delay,the adaptive particle swarm algorithm(APSO)is used to solve the system offsets.And the solution is simulated and compared with other methods.(4)Regional coordination control method based on dynamic subdivisionWith the goal of regional coordinated control,a dynamic subdivision method based on the improved Newman community algorithm is proposed.Taking the factors including the distance between the adjacent intersections,traffic volume,travel time,traffic flow discrete characteristic,signal cycle and road traffic flow density into consideration synthetically,the relevance of intersections is analyzed quantitatively.And relevancy degree model is established.Then the traditional Newman algorithm is improved and the intersections correlation degree is brought in to divide the regional network according to different traffic characteristics.On the basis of subdivision,a coordination control method of regional network is put forward.With the maximum throughout number of vehicles in the sub-region,the minimum of the intersection delay and stops as the optimized goals,a two-layer planning model is established to solve the public signal cycle of the sub-region and the green times of the intersections in the sub-region.Then offsets between adjacent intersections are obtained by the minimum delay arterial coordination control model considering vehicle arrival and dissipation process.Finally,the performance of the signal scheme proposed in this paper is evaluated and compared with other methods.