Understanding congestion in space-time domain using performance measures is essential prior to suggesting improvement schemes to reduce congestion. With technological advances like Global Positioning System (GPS), many metropolitan planning organizations give more emphasis on travel time based performance measures to quantify congestion, than on traditional way of using volume-to-capacity (V/C) ratios. In India, often it may not be possible to use personal vehicles as probes for travel time data collection. However, the public transit buses fitted with GPS devices could be used as cheap and effective probes to estimate the congestion status of other types of vehicles in the stream. The present study is an attempt in this direction. Two bus transit routes in Chennai, India were considered as case studies in order to cover the wide range of geometric and traffic conditions on urban arterials. GPS-fitted buses on these routes were used as probes in congestion quantification. As the dwell time at bus stops is a unique characteristic of transit buses when compared to other vehicles in the stream, a methodology has been proposed to find the dwell times including acceleration and deceleration times based on the approaching and departing speeds at bus stops. Regression models were then developed to predict the Congestion Index (CI) for various types of vehicles using bus CI, weighted carriageway width and the presence or absence of signalized intersection as independent variables. The results are promising and could be considered for real-time display of congestion levels for Advanced Traveler Information System (ATIS) applications. © 2019 Budapest University of Technology and Economics. All rights reserved.