METHODOLOGY
To evaluate performance a full-scale signal simulator (ASC/3) was integrated with VISSIM and the following three scenarios were modeled.

1. Baseline conditions (Current signal timing without TSP.)
2. Active TSP (Current signal timing with active TSP that provides a 10-second maximum green extension and 5-second guaranteed green on other phases).
3. Adaptive TSP (Current signal timing with adaptive TSP that balances improvements in bus travel times against long delays on individual approaches to each intersection.)

The simulated network represented eight TSP intersections with timing and traffic turning movements derived from historical field data. Thirty (30) buses on 25 routes had 15 minute headways during peak periods. Estimated maximum queue lengths were less than 200 meters at each intersection, and bus detectors were placed 250 meters upstream from each intersection to allow TSP solvers enough time to calculate optimal TSP timing for adaptive control.

A statistical study (t-test) was conducted to assess the significance of improvement by the proposed adaptive TSP.

FINDINGS

The results indicated that adaptive TSP can significantly outperform active TSP by dynamically managing control delay on bus approaches and non-bus approaches.

• Adaptive TSP can limit bus delay while maintaining an acceptable level of service to all traffic on all approaches.