Citation

Sohrabi S, Lord D. ITE J. 2019; 89(12): 29-36.

Copyright

(Copyright © 2019, Institute of Transportation Engineers)

DOI

unavailable

PMID

unavailable

Abstract

In this paper, a Bayesian hierarchical spatial model was developed for evaluating the effect of red light cameras (RLC) on injury crashes. The proposed model provides the capability of encountering unobserved heterogeneity in crashes and spatial dependency between intersections as well as capturing the spillover effect of RLC in the network. The model was developed using the data from Chicago intersections. Among the various land-use and intersection characteristics, the crash frequency was associated with the annual average daily traffic per lane, the number of lanes of intersections' approaches, the presence of a divided median on the minor approach. The results shed further light on improving the impact of RLCs on intersection safety by reducing the risk of injury crashes by 6 percent (all collision types). In addition, 2 percent fewer crashes are expected at intersections within 1 km network distance to the RLC location. From a practical standpoint, the proposed model for analyzing the RLC performance can result in a reliable assessment of the program. Also, results of this study can help previous attempts to investigate the economic feasibility of RLC programs and the allocation of RLCs in the network to achieve the highest efficiency.The authors defined an optimization problem using the captured RLC impacts, including the spillover effect, with the aim of maximizing the efficiency of RLC program in Chicago.

RESULTS show that the system performance can be improved by 13 percent after optimal camera allocation. More than 25 percent of the cameras need to be relocated to ensure maximum efficiency.

Language: en

Keywords

Cameras; Highway safety; Injuries; Intersections; Red light running; Spatial analysis; Traffic crashes; Bayes' theorem; Chicago (Illinois)