India Welcome MADRaS, The Open Sourced Driving Simulator Invented By Two IIT Graduates

Two IIT graduates have introduced a simulator - ‘MADRaS’ - which is considered the most advanced autonomous driving simulator at present. What inside MADRaS helps it get such high appreciation?

Autonomous driving has shown a variety of significant potentials, but there have been many obstacles like the quality of Indian roads or traffic jams. Realizing all difficulties waiting ahead, the two IIT graduates have developed a simulator that works for many vehicles at a time. They are Abhishek Naik from IIT-Madras and Anirban Santana from IIT-Kharagpur.

Their invention – MADRaS – stands for ‘Multi-Agent Driving Simulator’ and it is famous for the capability to test algorithms to drive several cars automatically at a time, even when they are under heterogeneous traffic conditions. When testing this simulator, the developers create a real-world traffic scenario to check cars’ reactions.

According to the developers - Abhishek Naik and Anirban Santana, MADRaS is a kind of multi-vehicle extension of Torcs which is usually applied for researches on autonomous driving. And hence, up to now, it has become the first open source multi-agent driving simulator not only in India but also in the world.

In addition to MADRaS, there have been many other simulators developed to support researches on autonomous driving including AirSim, Deepdrive, and Carla. However, all of the three simulators can only be applicable in one vehicle, and as a result, instead of controlling all cars simultaneously, controlling car’s behavior will limit in just one car. The other cars will keep running according to the pre-set regime. Along with the release of MADRaS, the restriction available in the abovementioned simulators will completely disappear.

MADRaS helps researchers to simulate several traffic scenarios like when there are a lot of cars moving fast or slowing down on the streets. Moreover, with the advantages of an open-source extension, this simulator is also highly appreciation due to its comprehensive coverage that allows other developers to create their own autonomous driving apps based on MADRaS.