Intelligent control method of signalized intersection based on dynamic phase optimization

WEI Liying; DING Canrong; AI Ziyan; GUAN Shuwen

doi:10.11860/j.issn.1673-0291.20220145

Abstract：In response to the current practice of using fixed phase and timing schemes in most signalized intersections in China,this study focuses on dynamic variable phase strategies and optimization of in-tersection signal timing.Firstly,the study calculates the left-turn traffic capacity based on critical gap theory,compares it with actual left-turn traffic flow,and selects an appropriate phase control scheme from the phase library.Subsequently,a multi-objective signal timing optimization model for signal tim-ing optimization is constructed,with average vehicle delay,average pedestrian delay,and stopping rate as optimization criteria.The model incorporates two sets of weight systems and utilizes the fast elitist non-dominated sorting genetic algorithm to identify the Pareto optimal solutions.The using variation coefficient method is utilized to obtain the first set of weight by using variation coefficient method;and the index coefficient method is used to obtain the second set of weight.The genetic algo-rithm is then employed to solve the multi-objective model as a single-objective model under both weight sets to derive the optimization scheme.Comparative analysis of the optimization results is con-ducted,followed by simulation using the Vissim software to validate the model.The results show that after optimizing the signal control schemes,the two-phase control scheme during peak hours at the studied intersection can reduce average motor vehicle delay and average pedestrian delay by over 50%and the stopping rate by more than 20%.This supports the feasibility of dynamic phase control strate-gies and signal timing optimization methods.