Manipulator Control using Federated Deep Reinforcement Learning

Overview:

I have done this project under the guidance of Dr. Nippun Kumaar A.A. and it explores the use of Deep Reinforcement Learning (DRL) for controlling robotic manipulators, enabling the arm to move from a random initial position to a defined target point. To address growing concerns around data privacy in robotic environments, the training process is extended using a Federated Learning (FL) framework, allowing multiple clients to collaborate without sharing raw data.

Tools used:

1. Pytorch Framework

2. Gazebo

3. ROS 2

4. Arduino

View paper: Paper

Video Demonstration: