FEDERATED DEEP REINFORCEMENT LEARNING FOR MOBILE ROBOT NAVIGATION
FEDERATED DEEP REINFORCEMENT LEARNING FOR MOBILE ROBOT NAVIGATION
I have done this project under the guidance of Dr. Nippun Kumaar A.A. and Dr. Amudha J. This project explores the navigation of a mobile robot in an unknown environment using Deep Q-Learning, enhanced with a federated learning framework to address data privacy concerns. The robot is trained to autonomously reach a charging point within a greenhouse setup, achieving a 100% success rate in both traditional and federated DRL models. By incorporating federated learning, the system ensures privacy-preserving training while still maintaining strong performance, making it suitable for distributed robotic applications.
Pytorch Framework
Webots
View Code: Code
View Paper: Paper