Theoretically, adaptive control strategies (ACS) have the ability to reduce a performance index (like delay, emission etc.) at an intersection when compared with pre-timed traffic signal control strategies. However, theoretical levels of benefit are not always achieved when these systems are implemented under real-world conditions. Errors due to detection and estimation procedures may have different effects on networks with different characteristics like geometry, traffic demand and number of phases in signaling scheme. Hence, it is beneficial to know how much improvement can be expected when these strategies are implemented on networks with different characteristics. This study compares the performance of three adaptive control strategies by using microscopic simulation tool PARAMICS. Prototypes for reactive (SCOOT-like), Case-based reactive (SCATS-like) and proactive / predictive (OPAC-like) algorithms, each using a different control logic to control signal timings, were developed in Paramics Programmer. These prototypes were tested on various well-calibrated NJ state highway intersections. It was observed that reduction in travel time achieved using these various adaptive control strategies peaked at different volume-to-capacity ratios for networks with different characteristics. The paper also identifies salient features of these distinct adaptive control strategies in terms of benefits and concludes that the benefits from these strategies tend to decrease with higher demand on cross streets at higher volume-to-capacity ratios. Jug-handle turns or closely spaced intersections also affect benefits that can be achieved using these algorithms. With higher number of phases and highly varying traffic demand, reactive algorithms are observed to be slow to respond, decreasing the level of benefits.
Abstract