Travel Time Estimation for Emergency Vehicles Using Bayesian Data Augmentation

• Activity Number: 295
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Bradford S. Westgate*+
• Companies: Cornell University
• Keywords: Travel times ; EMS ; Ambulance ; Reversible-jump ; MCMC ; Gibbs sampling

Abstract

Estimates of ambulance travel times on road networks are required for effective emergency medical services (EMS) planning. We introduce a new method for estimating the distribution of travel times on each road segment in a city, using data from Global Positioning System (GPS) devices on the ambulances. Due to sparseness and error in the GPS data, the exact ambulance paths and travel times on each road segment are not known. To estimate the travel time distributions using this data, we must infer the path driven from the GPS data. This is called the "map-matching" problem. We consider the unknown paths and travel times to be missing data, and simultaneously estimate them and the road segment travel time parameters using Bayesian data augmentation. We evaluate our method as a solution to the map-matching problem and the travel time distribution problem on a small region of Toronto, using a simulation study and data from Toronto EMS. We compare mean travel time estimates from our method with estimates from a speed averaging method, representative of currently used techniques.