Reinforcement Learning & it's Application on Autonomous Driving

In this project we look at some of the basic building blocks of reinforcement learning and try to understanding and implement the constituent algorithms. We start out by talking about the various types of reinforcement learning problems that are out there in the world. We give a little mathematical background for each one of these problems. We then move onto to focusing more only on one sort of learning problem which is associated with real-life problems and implement a problem on the python interface.

​

In phase 2 of this project, having familiarized ourselves with the fundamentals of reinforcement learning. We fabricate our own problem statement and design an environment from scratch. We also assembled a holonomic robot with the help of a microcontroller, that can detect images, classify them and carry out basic motion commands. By the fag-end of the project we gained a deep understanding of the pre-requisites for an R.L environment. Fabricating a rudimentary and fundamentally sound environment enabled us to think and poten- tially implement prospects involving a much higher level of complexity.