Critical evaluation of PASSER IV: a progression-based network signal timing program

Abstract

This thesis documents the evaluations performed on PASSER IV-94 Version 2.0 (PASSER IV) a new progression-based network signal timing program developed by Texas Transportation Institute. The signal timing parameters, namely cycle lengths and green times of PASSER IV, were compared with those of PASSER 11-90 Version 2.0 (PASSER 11) for two arterial streets. PASSER IV was compared with PASSER 11 for progression parameters, especially bandwidth, speed, and intersection spacings for four arterial streets. Delay estimates of PASSER IV were compared with PASSER II and TRANSYT7F Version 7.0 (TRANSYTI-7F) delay estimates for an arterial street. The results showed that PASSER IV and PASSER II assigned nearly similar green times for undersaturated arterials for both left and through movements. For saturated arterials, there is a significant difference in the allocation of green times, more so for left turning movements than through movements. The comparison studies for the progression parameters showed that PASSER IV in most cases had a larger bandwidth than PASSER 11. For cycle lengths where PASSER 11 had more bandwidth, PASSER IV coded with PASSER 11 green splits gave equal or more bandwidth. For optimized phasing sequence and speed variation between links, PASSER IV showed larger bandwidths in most cases than PASSER 11. Even though the logic for delay estimation in PASSER IV and PASSER 11 are similar, both programs gave very different results. PASSER II showing less delay than PASSER IV. TRANSYT-7F showed less delay than PASSER 11 or PASSER IV in case of total intersection delay especially for runs having larger cycle lengths. PASSER IV proved to be an excellent tool for arterial progression when compared to PASSER 11. The green splits of both programs are very similar and the differences in them are attributed to the fact that PASSER 11 has a green split optimization for delay while PASSER IV assigns them based on Webster's Equation. The differences in delay between both the programs should be further examined, based on the results of this work.