36 students across 8 universities build a 30-40ms V2V warning system using computer vision and Xbee radios — now under consideration for 5G pilot deployment.
India's roads claim over 200,000 lives per year — one of the highest road fatality rates in the world. Traffic accidents are the leading cause of death for young people aged 19 to 25, and a significant proportion involve large vehicles striking pedestrians and cyclists. The challenge of making Indian roads safer is immense, but a multi-year student engineering program demonstrated that locally-built technology could make a meaningful difference.
The program began with four student teams focused on using video analytics to identify and alert drivers to potential collisions. Using computer vision algorithms, the teams built systems that could analyze camera feeds from vehicles and classify driving scenarios as high or low collision risk — giving drivers a warning signal before impact occurs.
The Year 1 work focused particularly on the scenarios most common on Indian roads: interactions between large vehicles and vulnerable road users like cyclists and pedestrians at intersections and on highways.
In the second year, a student team focused on preventing chain-reaction accidents — when a collision between two vehicles causes a cascade of additional impacts as following vehicles fail to stop in time. Their solution: a vehicle-to-vehicle (V2V) communication system built on Xbee radio modules — off-the-shelf IoT hardware available at very low cost.
The system works as a real-time broadcast network:
The multi-year program involved 36 students across 8 universities. A working testbed was deployed at IIT Delhi's campus. The project earned a poster presentation at ACM MobiCom — one of the top academic conferences in mobile computing. The technology was identified as a potential use case for 5G infrastructure deployment.
The U.S. National Highway Traffic Safety Administration has identified V2V communication as one of the most promising technologies for reducing traffic fatalities. This student project demonstrates the technology is accessible to developing-country researchers using low-cost commodity hardware.