This project is about training a neural network to drive autonomously in an environment when planning its trajectory based on the road form (i.e. cones to mark the road).
As a part of the Formula project of the Technion, this is the algorithmic part that should be placed on the car’s computer to be able to take decisions in real time.
This already trained network should be re-trained in the exact environment in which the car will be finally tested or in another environment “close” as possible to it.