This study used the RHODES/BUSBAND model (Real-time Hierarchical, Optimized, Distributed, and Effective System) to simulate an intelligent bus-priority system in the suburbs of Seattle, WA, and Tucson, AZ. Each model simulated traffic activity along an arterial route having several intersections. The objective of the simulation was to optimize traffic-signal-phasing for a model network where real-time information on passenger counts and schedule adherence was available.

Initially, the decision logic used in the RHODES/BUSBAND model was applied to the CORSIM simulation model to generate baseline traffic conditions and to estimate the most practical locations for traffic sensor (loop detector) installation. Computer generated animations were used to confirm the baseline results.

Adaptive-signal-control was applied to the following baseline signal control systems in each study area:

• Fixed-time signals (pre-set times).
• Actuated signals (sensors triggered each way).
• Semi-actuated signals (sensors triggered for small road access to larger road).

The RHODES/BUSBAND simulation used "phase-constrained" or "weighted-bus" priority logic to control intersection phasing. The "phase-constrained" measure attempted to provide green signals for approaching buses. The "weighted-bus" measure gave extra priority to buses behind schedule with many passengers. In order to balance and optimize network travel time for all vehicles the weight of a passenger vehicle increased as its wait time increased at an individual intersection.

Each simulation model generated bus travel times, car travel times, bus delays at intersections, car delays at intersections, and total person delays. Buses were started at the main-street of each simulation model and then were subject to one of the following network designs:

• Semi-actuated control signals.
• No bus priority.
• Bus priority.

SIMULATION RESULTS

Adaptive-signal-control (ASC) with or without bus priority significantly increased average travel speeds and decreased total traffic delay. In addition, the average and variance of bus delays decreased while having little effect on other traffic.

The following table represents the average percent reduction in travel time observed when Phase-Constrained adaptive signal control systems were applied to a network of semi-actuated control intersections.

ALL VEHICLES	ASC Without Bus Priority	ASC With Bus Priority
Main Street Links	-7.94 %	-8.47%
Cross Street Links	-24.33%	-21.22%

The following table represents the average percent reduction in travel time observed when Weighted-Bus adaptive signal control systems were applied to a network of semi-actuated control intersections.

ALL VEHICLES	ASC Without Bus Priority	ASC With Bus Priority
Link with high cross-street volume	-0.43%	-5.20%
Link with low cross-street volume	-0.27%	-2.58%

The following table represents the average percent reduction in total person-delay observed when phase-constrained adaptive signal control systems (ASC) were applied to a network of semi-actuated control intersections.

ALL PASSENGERS	ASC Without Bus Priority	ASC With Bus Priority
Main Street Links	-18.19 %	-18.51%
Cross Street Links	-36.39%	-28.37%