In this project we leverage Scalable and decentralized Reinforcement Learning Algorithms
to improve cooperative decision making among autonomous agents in urban traffic scenarios.
In this project we propose a novel variable step RRT* algorithm with a dynamically time varying cost function to plan the best possible real-time trajectory amidst challenges of Wind and moving obstacles.
In this project we evaluate CNNs for prohibited item detection using real and synthetically composed imagery created using object extraction , region filling and threat object insertion.