News & Updates

Indo-US Workshop on
"Big Data Analysis for Transportation Engineering Systems " December 05-06, 2016

Workshop on Urban Freight Transport: A Global Perspective June 24-25th, 2013

"RESEARCH SYMPOSIUM 2012", Delhi, Dec 5th, 2012

National Conference on
July 13-14, 2012

Indo-US workshop,
Feb 11-13,2010

MoU Signing at New Delhi
June 21, 2010
September 19 -24, 2011

Real Time Bus Arrival Time Prediction System

Summary of Real Time Bus Arrival Time Prediction System
Intelligent transportation systems (ITS) are transportation services and technologies aimed at enhancing the efficiency, safety, reliability and eco-sustenance of transportation systems, without constructing new infrastructure. An important aspect of ITS is to advance public transportation to make it more attractive than private transport. India has traditionally boasted an extensive public transportation system, being the second largest producer of buses, accounting for 16 percent of world's total bus production. However, the share of public transportation in Indian cities has been on a steady decline over the last few decades due to, among other reasons, poor management of services. One of the major factors responsible for the success or failure of any public transport service is its reliability. One way of improving the reliability is to provide the passengers with accurate and reliable information regarding the service. This report describes a study performed on this important aspect, viz. bus arrival time prediction, a key field covered under the wide umbrella of ITS. This study is a part of a larger investigation and this report presents preliminary observations. During this phase of research, analyses were first performed using the GPS data collected manually from buses running in route numbers 21L and 21G in Chennai, India. Once the performance was found to be acceptable, online data integration was carried out for route numbers 5C and 19B. In order to have a real time automated application, it was required to develop an automated data collection filtration and analysis system without manual intervention. Numerous issues were identified and addressed during real-time implementation of the model. Issues such as effect of traffic jam, overtaking among the buses on the same route, bus breakdown, abrupt changes in bus routes, etc. which are specific to Indian conditions were taken into consideration. A model based prediction scheme was used and the Kalman Filtering Technique was adapted to estimate and predict the travel time of buses. Different modifications to the system were attempted for improving the performance and those that were improving the performance were incorporated into the model. The study also developed a prototype of the complete system integrating with the information dissemination units. Prototypes for three different information dissemination modes were included, viz. bus stop VMS display, Kiosk display at bus stops and web based application. An advanced kiosk display system was developed considering specific requirements of users. Such a system has advantages of cost efficiency and provides greater information in a cognitively ergonomic format to aid the commuter’s decision making skills. Although the application developed covers all aspects of real-time implementation, there is still scope for improvement. Search for better algorithms for more accurate prediction is an open ended problem. In the present study due to limitations of the KFT based models, the data from the test vehicle was not used in the algorithm and was kept for validation purpose alone. New algorithms that would use real-time data from the test vehicle are being explored. Modification of prediction algorithm will be an ongoing process along with evaluations. Transferability and scalability of the system need to be taken into account.

---> Click here to see APTS_Finalreport.PDF