Citation

Alecsandru C, Rouhieh B, Huang TKC. J. Transp. Saf. Secur. 2010; 2(4): 336-351.

Copyright

(Copyright © 2010, Southeastern Transportation Center, and Beijing Jiaotong University, Publisher Informa - Taylor and Francis Group)

DOI

10.1080/19439962.2010.522301

PMID

unavailable

Abstract

In designing intersection signal timing plans transportation professionals have to account for two main objectives that often times are antagonistic: to ensure good flow of traffic and to maintain a high level of safety for all road users. Although the two objectives do not necessarily exclude one another, identifying the compromise that provides the best traffic and safety conditions for all road users is not a straightforward exercise. In this study, the authors propose a methodology that can be used to determine the best signal timing of urban intersections by reaching a desired equilibrium between the two objectives. The methodology is using a combined delay-safety (DS) performance measure in an artificial intelligence decision-making framework. A case study of an urban arterial with a newly built bicycle path in downtown Montreal, Quebec, was investigated using a microscopic traffic simulator, VISSIM. A multilayer perceptron (MLP) neural-network uses several traffic flow parameters as input information to identify, out of three possible configurations (i.e., independent signals, coordinated for automobile progression, and coordinated for bicycle progression) what type of signal timing plan yields the best tradeoff between automobile delay and safety of nonmotorized users. Based on several levels of input flows from real-world and simulated traffic data a large pool of possible input/output combinations was used to train and test the MLP neural network with two hidden layers. It was found that for 99.8% of the tested cases the neural network identifies correctly the configuration of signal timing plan that yields the lowest DS value.