Adaptive signal controls are currently the subject of much attention around the world due to the potential flexibility they offer. The increased sophistication of their logic provides greater flexibility in signal control, but also makes the evaluation of their performance more difficult. While there are many techniques used to estimate delays for fixed time signals, none of these techniques can be used to obtain reliable delay estimates for adaptive signal control systems. In this paper, an approach to estimate vehicular delay at interesections controlled by adaptive signal systems is introduced. Vehicle-by-vehicle arrival times are collected on the upstream approach to the intersection. An event-based simulation algorithm uses these arrival data, together with signal timing data, to track the vehicle movements throughout the intersection. The algorithm estimates the time when the vehicle stops at the intersection, if it stops, as well as the time it departs the intersection. Other traffic characteristics, such as the number of stopped/non-stopped vehicles and the number of vehicles in the queue, are also estimated. The algorithm was developed and tested using data from South Lyon, Michigan. The results showed that the difference between estimated and observed values of average delay averaged 1.76 sec/vehicle. The difference between observed and estimated departure time was less than 5 seconds averaging 2.14 sec with 38% of the vehicles having no error in departure time estimates. The estimates of number of stopped/non-stopped vehicles had no error in 74% of the cycles. Delay estimates obtained from the algorithm were more accurate than those estimated by the HCM, especially when the degree of saturation exceeded 0.8. Based on these results, it can be concluded that the proposed algorithm can be effectively used to estimate vehicular delay, queue length, and stopped/non-stopped vehicles at intersections.
Samenvatting