Project: Self Driving Car using Neural Networks

Suggestions welcome :)

Introduction

This project uses deep neural networks to learn how to drive a car. A car can go forward, backward, left and right. The goal is to teach the car to drive itself amongst the traffic. We simulate traffic by placing other cars on the road. The car will learn to avoid collisions with other cars and stay on the road.

Project Structure

There are multiple JS Scripts that do different jobs, plus there is a HTML and CSS file.

• index.html: ok.
• style.css: pain.
• car.js: Object constructor and methods for each car
• road.js: Object constructor and methods for the road.
• awareness.js: Creates the sensors in front of the car to detect other cars or road boundary.
• network.js: Initialise the neural network.
• calculation.js: some math stuff I need to detect collisons.

Demo:

Check it out at: https://harmya.github.io/SelfDrivingCar/