Research On Optimization Of Single Intersection Actuated Control Under Stochastic Conditions

There are two control schemes at signal control intersection:fixed time control andactuated control. According to actual traffic informations from detectors actuated control canadjust traffic signal dynamically, and it overcomes the defect of fixed time control. A welldesigned signal timing scheme of actuated control not only can effectively improve the trafficcapacity, reduce the traffic accidents frequency, but also can reduce traffic pollution to someextent. Actuated control parameters and control strategies decides the finally control effect. Inthis paper, by considering the stochastic conditions of traffic flow, the minimum green time，maximum green time and the control logic model are proposed.Firstly, the paper analyzes the existing determination of timing parameters in theactuated control system. During calculating the minimum green time and maximum greentime, the stochastic conditions of traffic flow was not fully considered, it assumed that thearrival rate obeied uniform distribution which were not agree with practice situation. Thatresults in certain limitations to reach the expected effect in application process. Secondly,considering the stochastic characteristics, the minimum green time optimization model isestablished where the average delay is used as the optimization objective under differentarrival distribution, subject to the constraint of the safety concern and cycle length. UsingLagrange analytical method, minimum green time model is solved, and the optimal value ofeach phase minimum green time is obtained by approaching the objective function withstochastic simulation method. Then the relationship between the green extension time andheadway of stochastic traffic is analyzed. Considering the probability of variouscircumstances extension, the maximum green time optimization model is established.According to analyzing the dynamic phase sequence in fixed timing control, phase sequenceoptimization method is proposed based on queue length of critical phase approach. In order toavoid of oversaturated phase which can not able to get right-of-way in time, the longest queuelength phase is selected as the next phase, so it can ensure the effective use of green time.Finally, using the actuated module of VISSIM to simulate the proposed model. Comparedwith the traditional actuated control method, a simulation example of a signalized intersectionis presented to experiment and analysis the model. The results show that the proposed modelhave different degrees of improvement under the condition of different saturation of eachevaluation when the vehicle arrival rate obeies the Binomial distribution and Poissondistribution.