“Travel Time Reliability Prediction: Mapping Services to Ambulance Fleets” by Dr. Dawn B. WOODARD
Dr. Dawn B. WOODARD
School of Operations Research and Information Engineering
Mapping services like Google Maps and Waze provide predictions of travel time for arbitrary routes in a road network; however, there can be considerable uncertainty in those predictions due for example to unknown timing of traffic signals and uncertainties in traffic congestion. Probabilistic forecasts of travel time improve over the deterministic predictions from mapping services, by accounting for this uncertainty. Such probabilistic forecasts of travel time can be used for risk-averse routing, for reporting travel time reliability to a user, or as a component of fleet vehicle decision-support systems. I will present methods to predict the probability distribution of travel time on an arbitrary route in a road network at an arbitrary future time. Our methods give informed predictions for parts of the road network with little data, capture weekly cycles in congestion levels, and are computationally efficient even for very large road networks and datasets. I highlight two important cases: (1) commercial mapping services, and (2) emergency vehicle fleet management, and present methods appropriate to each case. For mapping services the most relevant source of information is GPS data from mobile phones, while for emergency vehicles it is automatic vehicle location data from fleet vehicles. Our methods are applied to (1) large volumes of mobile phone data from the Seattle metropolitan region, and (2) the Toronto Paramedic Services ambulance fleet.