Author : Prof. Arvind Nair
Abstract: Autonomous navigation remains a cornerstone in the advancement of intelligent robotic systems. The complexity of real-world environments presents significant challenges for traditional navigation algorithms. This paper explores the application of Deep Reinforcement Learning (DRL) for path planning and navigation in autonomous systems. DRL combines the perception capabilities of deep learning with the decision-making prowess of reinforcement learning, enabling robots to learn optimal navigation strategies in dynamic and unknown environments. Key architectures such as Deep Q-Networks, Actor-Critic models, and Proximal Policy Optimization are discussed in detail. Performance metrics, environment setups, and real-world case studies are included to illustrate the practical impact and limitations of DRL in autonomous navigation. This study concludes by highlighting future research directions and the potential of DRL to transform path planning in robotics.
Keywords: Autonomous Navigation, Path Planning, Deep Reinforcement Learning, Deep Q-Networks, Actor-Critic, Proximal Policy Optimization, Robotics.
Full Issue
| View or download the full issue | PDF 38-41 |