In complex dynamic environments, robotic arms face multiple challenges such as real-time environmental changes, high-dimensional state spaces, and strong uncertainties. Trajectory planning tasks based on deep reinforcement learning (DRL) suffer from difficulties in acquiring human expert strategies, low experience utilization (leading to slow convergence), and unreasonable reward function design. To address these issues, this paper designs a neural network-based expert-guided triple experience replay mechanism (NETM) and proposes an improved reward function adapted to dynamic environments. This replay mechanism integrates imitation learning’s fast data fitting with DRL’s self-optimization to expand limited expert demonstrations and algorithm-generated successes into optimized expert experiences. Experimental results show the expanded expert experience accelerates convergence: in dynamic scenarios, NETM boosts accuracy by over 30% and safe rate by 2.28% compared to baseline algorithms.

Deep reinforcement learning (DRL) has become pivotal in robotic control for its superior learning and decision-making capabilities. This study focuses on compliant control of rehabilitation massage robots under dynamic uncertainty, proposing a method combining the SAC algorithm and PID feedforward compensation to enhance adaptability, robustness, and self-optimization in complex scenarios. For periodically dynamic simulation planes, the reward function is optimized to boost learning efficiency. To mitigate force errors from environmental uncertainty, PID feedforward compensation adjusts the robotic arm’s end-position deviation. Simulation experiments validate constant force tracking during massage, achieving stable control.