VividhaVahana: Smartphone Based Vehicle Classification And Its Applications in Developing Regions

Developing region road traffic has a unique characteristic of high heterogeneity in vehicle types. In this paper, we describe VividhaVahana, a smartphone sensor based system to categorize road vehicles into four predominant categories: two-wheeler bikes, three-wheeler auto-rickshaws, four-wheeler cars and public transport like buses. Using a variety of sensor based features, our system is able to achieve above 90% classification accuracy, evaluated over 1500+ Km of driving data, on two urban road stretches in the Indian city of Delhi. We also apply VividhaVahana to empirically examine four representative transport applications, namely travel time estimation, driving behavior detection, traffic state classification and road surface monitoring. We show how each of these applications would benefit from a vehicle class specific analysis, compared to the vehicle agnostic analysis as has been done in the past. Our work gives useful insights on how such applications can be re-designed, to better fit developing region traffic characteristics and requirements.