IGVC

Supervisors: Prof. Mangal Kothari, Prof. Gaurav Pandey

Report Website Code

Intelligent Ground Vehicle Challenge is a competition organised by AUVSI which is aimed at developing prototype self driving cars. The competition pits teams from around the world against each other on arenas like the one shown below. Teams are supposed to develop different algorithms to autonomously traverse the arena using any method possible.

Our team began working for this competition in the December of 2016, and we took part in the 2018 edition of the competition, where we secured the 5th position in the desing challenge and overall we placed 12th internationally.

Our inital approach was to use a LiDAR and fuse its data with IMU and GPS to develop a global map. This global map was also to contain obstacles from the lanes. We were also segmenting out the lanes using different classical Image Processing Techniques and then adding to the grid map generated from the fused data before. The resultant grid map was then planned on, which gave us the final trajectory to track.

We have evolved much from then, you can check us out on Github or take a stroll on our website.

I will be updating this page to include more information when I am free.

While on the team I worked on various parts of the problem from Sensor Fusion, Computer Vision to overall system Integration. As of now I am not actively working on the team but as a mentor.