Space Travel Time Information System in Stockholm City:prototype system and algorithm developments using AVI measurement data

摘要

This paper illustrates our development on travel time information system and research on real- time travel time estimation and prediction algorithms using Automatic Vehicle Identification (AVl) data in the Space Travel Time Prediction (STTP) project，funded by Trafikkontoret Stock- holm Stad (TSS).To support the implementation of a real-time travel time information system in Stockholm using AVI data collected on urban roads, a preliminary travel time analysis plat- form was developed and several filtering algorithms and their modifications were implemented and evaluated for travel time estimation purposes in previous work. In the current development we have extended the previous experiment platform into a web-based system in which travel time data arc stored in data warehouses based on spatial and temporal dimensions, and real-time esti- oration and prediction are available through WWW internet. In addition, we have developed an adaptive algorithm for on-line travel time prediction based on time-series modeliug approaches. Especially, the Kalman Filter has been utilized for the implementation of a recursive optimal predictor, which can handle the noise and non-stationary of travel time measurements.