Potential effectiveness of signal optimization for various corridors in Michigan

Abstract

This study investigates the potential effectiveness of signal timing at 130 intersections located on five corridors in southeast Michigan. Effectiveness is measured in terms of reduced numbers of crashes after signal timing was introduced. Five years of Michigan crash data from 2001 through 2005 were used to count crashes both before and after signal timing. For each of the 130 intersections, crash sites were geographically located on maps using a spatial analysis software tool and summary crash statistics were provided for injury severity, time of day, day of week, and crash type. In total, there were 12,438 crashes on the 130 intersections. Approximately 80 percent of these crashes resulted in property damage only, and about half were rear-end type crashes. A before-after statistical model was developed to assess the effects of signal timing on the numbers of crashes after the intervention. Results are provided for each intersection. Overall, the corridors are ranked in the following order, based on greatest reductions in numbers of crashes after signal timing: Jefferson Avenue, Plymouth Road, Woodward Avenue, Ford Road, and Hall Road. Jefferson Avenue, Plymouth Road, and Woodward Avenue showed overall reductions in crashes after signal timing. Ford Road showed no change, and Hall Road showed an increase in crashes after the treatment. It was hypothesized that the increase in crashes on Hall Road might be due to increased traffic volumes on that corridor due to community and economic development after the signal timing started. Examination of available average daily traffic (ADT) counts before and after treatment did not support that hypothesis. Crash type and crash severity were investigated to determine if any shifts occurred in these distributions after signal timing. It was found that intersections that showed a significant reduction in numbers of crashes after signal timing had higher percentages of angle crashes and lower percentages of same direction crashes than intersections that showed no change in crashes after signal timing.