Making the Vehicle (Car) Navigate in Traffic by Avoiding Collisions Using Deep Learning

Abstract

Deep learning (DL) which is subset of Machine learning (ML) along with Transfer-learning (TL) can significantly help vehicles (car’s) to achieve human- level ability to safely navigate in unpredictable dynamic environments, but in order to develop an autonomous vehicle we need lots of training data and obtaining data from real world using a car is a complicated task as it is time consuming to set up everything, as it requires high end cameras along with humungous amount of time in gathering the data, further it’s not very safe and it’s less convenient, so let’s try developing an autonomous vehicle using a game which will be able to navigate safely by avoiding collisions as much as possible here we are using a game to build an autonomous vehicle be cause it’s very easy to setup everything, a gaming environment is much similar to that of a real-world environment.