Technology developments to improve safety and efficiency of transit bus operations

Frontal Collision Warning (FCW) and Transit Signal Priority (TSP) have received significant attention in recent years as promising means to improve safety and efficiency of transit bus operations. The purpose of an FCW system is to detect a potential collision and warn drivers with sufficient time to react under imminent crash situations. TSP is an operational strategy that facilitates bus movement through signalized intersections. This dissertation focuses on developing advanced technologies for target tracking, threat assessment, and bus arrival time prediction.;In target tracking, target maneuvers may cause divergence of tracking as the maneuver magnitudes are not known. To overcome this problem, we develop a target tracking scheme that treats the maneuver inputs as a bias sequence and estimates it in real-time. Compared with other tracking schemes, the developed scheme has two advantages. First, it responds to maneuvers faster while maintaining the same level of tracking quality. Second, it has the ability to adapt to a wide range of maneuver levels without changing the assumed level of process noise.;A successful threat assessment algorithm must promote a timely and appropriate driver response while balancing system effectiveness against false/nuisance alarm rate. Using real-world data, braking onset parameters by transit drivers are investigated. This study provides insight into when a bus driver would initiate braking and how the driver adjusts braking to control the vehicle motion. We present evidence to demonstrate that the Time-To-Collision (TTC) is the most consistent measure of braking behaviors, and that a TTC measure of 4 sec is a proper warning strategy.;Predicting bus arrival time is challenging because of the variations in traffic conditions. We develop an adaptive algorithm that utilizes the following two models: a historical model that provides predictions based on average traffic patterns, and a real-time model that produces predictions based on the current traffic conditions. By assigning weightings to these two models, the adaptive algorithm balances the uncertainty in downstream traffic and the knowledge of current surrounding traffic. As a result it provides a higher level of prediction accuracy. (Abstract shortened by UMI.)...