Autonomous Driving Using Machine Learning

Abstract

An autonomous car is a vehicle which is capable of operating i.e. driving on its own without human intervention. Until 2015, a fully autonomous car was just a dream. But due to rapid technological advances, the dream eventually became a reality. Autonomous vehicles which were able to perform just basic operations like moving forward or backwards with the help of a remote control are becoming fully autonomous by the day. The hideous looking autonomous car which had a ton load of sensors mounted on the roof is now becoming sleeker and nattier. Autonomous vehicles detect the surroundings using various techniques such as RADAR, LIDAR, GPS, odometer and computer vision. Advanced control systems interpret sensory information to identify appropriate navigation paths, as well as obstacles and relevant signage viz. wrong way, no right turn, no left turn, etc. Autonomous cars have control systems that are capable of analysing sensory data to distinguish between different cars on the road, which is very useful in planning a trajectory in which the car should maneuvre in order to avoid a collision. This paper focuses on autonomous driving and demonstrates how a car will be able to drive without any human intervention by processing the input data taken from the sensors and applying various machine learning techniques to it, and thus making the correct decisions.

Keywords: Deconvolution, Hough transform, Internet of things, Lane detection, Machine learning, Neural networks, Swarm intelligence.

View PDF