Estimating travel times using Wi-Fi sensor data

Determining travel time information from Wi-Fi (or Bluetooth) sensors is not trivial due to various (often technical) reasons. In this contribution, we focus on the problem of distinguishing travel time from the time people spend performing activities (e.g. fuelling the car, standing still to watch the scenery, buying a train ticket). More specically, we will consider pedestrian data collected during a large-scale event in the city of Amsterdam called SAIL, where visitors walk along a route while watching and visiting tallships, eating and drinking (https://www.sail.nl/EN-2015). In this specic type of application, travel time information is required to provide information about the delays due to crowding, while the time spent on performing activities does not reflect such crowding effects.In this contribution, we will present a novel statistical approach to estimate travel time distributions from data collected by Wi-Fi sensors.