Published: 12th August 2019
Is the traffic getting to your nerves? These IITans have a solution to solve the crisis
The given task for the student competition was to help the police predict traffic volume for given duration and climate conditions
Frustrated to drive to work pushing through all the traffic? The brightest minds from our IITs might just have a solution for the irritating traffic we get stuck in every day.
The IIT-Madras Alumni Association (IITMAA) on Monday announced the top 50 teams in its Machine Learning Hackathon Contest being held as part of its Annual Flagship Event 'Sangam,' which will deliver a strategy document for 'Reimagining India in 2030' to the Prime Minister's Office (PMO). And the scenario for the Hackathon was an Indian Smart City's Traffic Police Department has decided to use Machine Learning/Artificial Intelligence techniques to solve their traffic crisis and collected traffic volume patterns and climate conditions observed for over four years, a release from IIT-M here said.
The given task for the student competition was to help the police predict traffic volume for a given duration and climate conditions. As many as 3,266 teams participated in the Phase-1 online hackathon with the top teams making predictions with an accuracy of up to 99.97 per cent using regression models such as 'XGBoost,' supported by adept Feature Engineering techniques.
The top 50 teams of Phase-I were from IIT-Bombay, IIT-Delhi, IIT-Kharagpur, IIT-Roorkee and IIT-Madras, besides Indian Statistical Institute, National Institute of Technology, Jamshedpur, National Institute of Technology, Patna and Army Institute of Technology, Pune.
Krishnan Narayanan, Co-Chair, Sangam 2019 and President, itihaasa Research and Digital, said, "Such highly accurate AI/ML models and analysis can not only help Indian Traffic Governance Systems predict and effectively manage, the traffic volume for a defined time period but also help analyse a trend to observe the correlation between Air Pollution Index and Traffic Volume." The demographics of participants include 257 teams from Delhi NCR, 288 from Mumbai and Pune, 509 from Chennai and Hyderabad, 437 from Bengaluru, 103 from Kolkata, 50 from Coimbatore and 1,622 from other parts of the country.
In Phase-2, IITMAA will conduct a Hackathon onsite at the IIT-Madras campus on August 24 and 25. The Hackathon will provide the short-listed teams time series data on air quality and pollution recorded by IoT (Internet of Things) devices from 'Project Kaatru (Air)', a research project of Prof Raghunathan Rengaswamy, IIT Madras and supported by Robert Bosch Centre for Data Science and Artificial Intelligence, IIT-M.