Q-Learning for Driving on Infinite Road
Loading...
Date
2018-11-01
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The autonomous vehicles will consume large amounts of streaming data from nearby vehicles, infrastructure, and cloud to make decisions. The goal of the project is to adaptively change the velocity of a connected autonomous vehicle to minimize the fuel consumption of the vehicle. The vehicle will make real time decisions based on real-time information from V2X and V2V connectivity keeping in mind the constraints imposed by the powertrain of the vehicle. We aim at solving this problem using reinforcement learning algorithm. The ultimate goal of the project is to design and run very large scale reinforcement learning algorithm for devising fuel efficient driving strategies.
Description
Keywords
Reinforcement Learning, Autonomous Vehicle, Statistics