GSM-based positioning for public transportation commuters

Crowd-sourcing of information about various aspects of a road is an important mechanism in Intelligent Transportation Systems (ITS). In this work, we consider crowd-sourcing from public transportation commuters, information about the road, traffic, and the specific public transportation unit. Any crowd-sourced information has to be tagged with the information provider's location. Since GPS-based location determination is energy-expensive, we focus on GSM signal based location determination. A specific dimension we explore, not considered in prior work, is the use of GSM signal information from multiple commuters' phones. This consideration is triggered by the observation that a large set of commuters have (almost) the same physical location for large durations of time: when they share the same bus. We present an analysis of data collected from two different Indian cities: Mumbai and Chandigarh. We find that such combination of information can lower the median location error by a factor of 2-10.