Evaluation of detector placement strategies for first generation traffic responsive signal control

Abstract

Traffic responsive signal systems rely on system detector data to evaluate current traffic conditions and select a corresponding timing plan from a library of plans. One of the major decisions surrounding the use of traffic responsive signal systems involves locating system detectors within an arterial network. Simulation studies were undertaken using TRAF/NETSIA4 software to evaluate detector configurations for a traffic responsive signal control system on NASA Road I in Houston, Texas. Three different detector configurations were examined: the existing configuration used by the Texas Department of Transportation, a major intersection and 800 meter placement, and use of detectors only on the approaches to the critical intersection. Traffic responsive operations were simulated for three time periods to evaluate system performance under varying traffic conditions. Time-of-day operations for the same three periods were also simulated to provide a basis of comparison in evaluating the effectiveness of detector configurations. System performance was analyzed on the basis of total system delay and the number of detectors required by each detector configuration. It was determined that there were no statistical differences among the delay data while operating the different detector configurations and time-of-day mode. Given the second evaluation criterion, minimization of the number of system detectors, it was concluded that a well calibrated time-of-day plan (which required no system detectors) offered the best alternative. This conclusion was made on the basis that there was no statistical difference in the performance levels achieved by time-of-day operations and traffic responsive operations. In situations where the use of traffic responsive mode might be desired, it was concluded that operating in the traffic responsive mode using detectors at the critical intersection offered the best alternative of the three configurations considered. While this result was expected to hold true for relatively small arterial systems, its applicability to complex arterial networks with multiple critical intersections was not investigated. This research further suggested that the focus of traffic responsive control systems akin to the NASA Road I system be shifted to only respond to major directional traffic pattern changes and major shifts in cycle length. These observations were consistent with past research.