International Journal of Robotics and Autonomous Systems ISSN: 3107-7382 (Online)

Authors: Dr. Arvind Deshmukh, Ms. Meenal Rao

Abstract : Reinforcement learning (RL) has emerged as a powerful paradigm for enabling robots to learn complex tasks through interaction with their environment. In robotic assembly automation, where variability, precision, and adaptability are critical, RL provides a mechanism for machines to develop decision-making policies that optimize performance over time. This paper explores the implementation of reinforcement learning algorithms—such as Deep Q-Learning, Proximal Policy Optimization, and Actor-Critic methods—in industrial assembly lines. By simulating trial-and-error behavior, these methods enable robotic agents to perform complex assembly operations like insertion, alignment, and fastening with minimal human supervision. The study highlights the benefits, challenges, and future prospects of integrating RL into smart robotic systems for automated assembly.

Keywords: Reinforcement Learning, Robotic Assembly, Automation, Deep Q-Learning, Smart Manufacturing, Actor-Critic, PPO