ADAPTIVE CONTROL STRATEGIES FOR ROBOTIC MANIPULATORS USING DEEP REINFORCEMENT LEARNING

This research examines adaptive control strategies for robotic manipulators using Deep Reinforcement Learning (DRL) through a systematic literature review approach. The main focus of the study is the identification of commonly used DRL algorithms, implementation challenges, and the direction of developing DRL-based adaptive control systems. The study results show that algorithms such as DDPG, SAC, and PPO are effective in addressing the non-linear dynamics and uncertainties of robotic manipulators, both in simulation and real-world environments. However, there are significant challenges such as the need for large training data, the simulation-to-real gap, and the limitations in the interpretability of control policies. The integration of hybrid control strategies, the development of more sample-efficient algorithms, and the application of hierarchical and meta-reinforcement learning have been identified as promising future research directions. This study provides a foundation for the development of more flexible, efficient, and safe robotic control systems to support various industrial applications.