Determinants of Bus Riding Time Deviations: Relationship between Driving Patterns and Transit Performance

Abstract

Urban bus services are subject to high levels of uncertainty and disturbances. Methods to determine the timetable are designed to absorb variations in riding times between stops by allocating additional travel time. The propagation of service unreliability along the route could be restrained by drivers’ adjustment at stops and between stops. This paper analyzes the main determinants of bus riding times deviations based on automatic vehicle location (AVL) data from four trunk lines in Stockholm, Sweden. The analysis indicates that drivers can and do adjust their speeds in response to instantaneous real-time schedule adherence information, although these adjustments depend on the underlying control scheme: locations where the performance is measured. A model for bus riding time deviations was estimated with autoregressive effects, performance indicators, link characteristics, and trip attributes as the explanatory factors. The results can support the development of travel time prediction and real-time control strategies that take drivers’ response to operations into account. This highlights the importance of the human factor in designing control schemes and the corresponding transit performance evaluation.