Font Size: a A A

Research On The Optimization Of Signal Timing For City Single Intersection

Posted on:2019-07-02Degree:MasterType:Thesis
Country:ChinaCandidate:M ZhangFull Text:PDF
GTID:2382330548456908Subject:Transportation engineering
Abstract/Summary:PDF Full Text Request
Urban traffic congestion has become a common phenomenon during the peak hours in the morning and at night.As the key node of urban traffic,the operation efficiency of intersections plays an important role in optimizing the performance of road network and alleviating urban traffic congestion.Reasonable signal timing plan can greatly improve the operation efficiency of intersections.The existing signal control systems and researches focus on the arrival pattern and queue of vehicles in the intersection.When the intersection is saturated and oversaturated,these studies only focus on the entrance lanes which are saturated and oversaturated,ignoring the entrance lanes which are unsaturated,which can not realize the optimal allocation of space and time resources of intersections.To solve this problem,this paper presents an optimization method of traffic signal timing based on the value evaluation,considering the queuing condition of all entrance lanes in the intersection.According to the arrival and release characteristics of vehicles,this paper models the vehicle queuing process of entrance lanes in the intersection.And this paper establishes the model of intersection performance index,for example the vehicle waiting time ?the number of vehicles through the intersection and so on.Because of the large search space in the process of signal timing,which may lead to dimension curse,we establish an intersection signal timing optimization model referring to the learning process of Q-learning.First of all,based on the release efficiency of vehicles in the intersection,a reward function is constructed to evaluate the released flow and green time.And then,the value evaluation function is constructed based on the queuing vehicles and capacity of entrance lanes to evaluate the subsequent state of intersection,thus reducing the depth of the search tree.With the constraint of green time,we establish an intersection signal timing optimization model based on the value evaluation,combined with the reward function and value evaluation function.Finally,we select the intersection of HePing road and ShanDa road for example in JiNan to simulate,proving the validity of the presented method.The proposed method is compared with WEBSTER method and the existing signal timing scheme of intersection by the VISSIM simulation software.The results show that when the intersection is unsaturated,there are more vehicles through the intersection in a certain time via the method presented in this paper,and the waiting time of vehicles is also reduced.When the intersection is saturated or oversaturated,the WEBSTER method fails.Compared with the existing signal timing scheme of intersection,there are more vehicles through the intersection by the method presented in this paper,and the waiting time of vehicles is reduced.In addition,the method proposed in this paper can postpone the overflow of the entrance lanes of intersections,which can alleviate the traffic jam to a certain extent.
Keywords/Search Tags:Traffic congestion, Single intersection, Q-learning, Reward function, Value evaluation function
PDF Full Text Request
Related items