Font Size: a A A

Study On Dynamic All-red Control Method At Urban Signalized Intersections

Posted on:2017-03-25Degree:MasterType:Thesis
Country:ChinaCandidate:H J LiFull Text:PDF
GTID:2272330509956946Subject:Transportation planning and management
Abstract/Summary:PDF Full Text Request
As important nodes of city road network, traffic accidents often occur at signalized intersections. When the yellow-onset, drivers near the intersection would make a decision of braking to stop or continuing to travel, some of which continue to travel will go through during the yellow or red light and cannot go through safely, may collide with the first vehicle of next phase. Therefore, on the basis of observation data, to protect the vehicle’s driving safety and improve the safety level at signalized intersections, this paper analyzed vehicle’s traveling behaviors after the yellow-onset, built travel time prediction model, and researched the dynamic all-red control method.This paper selected signalized intersections in Harbin as observation sites, developed and recorded video observation scheme. At the same time, since the last-to-go vehicles and first-to-stop vehicles are most likely to make wrong decisions when the yellow-onset, and the last-to-go vehicles are latest to pass, so this paper extracted vehicles’ traveling parameters and road traffic conditions data of last-to-go vehicles and first-to-stop vehicles when the yellow-onset.Firstly, based on the observation data, this paper selected driving behavior factors, and solved the probability that drivers chose to continue to travel with binary logistic regression model, and realized the determination of vehicles continuing to travel after the yellow-onset. Secondly, for the last-to-go vehicles at each lane among vehicles continuing to travel, namely last vehicles of last phase, this paper used BP neural network and kinematic equations, solved exiting time of the last vehicle and arriving time of the first vehicle, proposed the setting condition and solving method of dynamic all-red. Finally, this paper designed the basic framework and control flow of dynamic all-red control method, and using one of the observation intersections data, evaluated the safety benefit and delay cost dynamic all-red control method, verified its expected performance.On the one hand, this paper’s research results effectively protect light-running vehicles, make them pass the intersection safely, avoid traffic accidents and improve the safety level of signalized intersections; on the other hand, with respect to the intersections with static all-red, can minimize the intersection delay under the premise of safety.
Keywords/Search Tags:signalized intersection, dynamic all-red control method, binary logistic regression, BP neural network, performance evaluation
PDF Full Text Request
Related items