Incorporating the Effects of Signal Transitions in the Selection of Timing Plans in Traffic Responsive Signal Systems (SWUTC/98/465080-1)

As part of this research project, a computer program, SIGTRAN, was developed to estimate the impacts of the transition phase when changing timing plans. The program uses the same procedures as PASSER II-90 to compute the estimated delay due to transition. Using the SIGTRAN program, a proposed algorithm for selecting new timing plans in a traffic responsive mode was developed. The algorithm compares the delay associated with retaining the old traffic signal timing to the delay associated the implementing a new timing plan plus the delay caused by transitioning to the new timing plan. If the delay for the old timing plan exceeded the delay estimates for the new timing plan plus the delay accrued during transition, then the new timing plan was adopted; otherwise, the old timing plan was retained. The new algorithm was tested using data from NASA Road 1 in Houston, Texas. It was found that the traffic responsive timing plans developed using the algorithm produced substantial delay savings in the A.M. Peak and Noon periods over operating the signals in a time-of-day mode. Less significant delay savings were generated during the A.M. Off-Peak P.M. Off-Peak and P.M. Peak periods. During these periods, either traffic volumes did not change significantly or the differences in the timing plans were so slight that most of the benefits of changing timing plans were offset by the increase in delays caused by transitioning to the new timing plans.