Abstract—Machinery system vibration control (e.g., aerospace, automotive, and robots) requires adaptive control techniques to address nonlinear dynamics and environmental uncertainty. The conventional approaches of Proportional-Integral-Derivative (PID) and Linear Quadratic Regulator (LQR) controllers are typically non-adaptive in nature for changing operating conditions. A hybrid approach is proposed in this paper for enhanced real-time active vibration damping. An innovative technique combining Deep Deterministic Policy Gradient (DDPG), a Reinforcement Learning (RL) algorithm, with fuzzy logic is developed. The fuzzy system tracks uncertainties in sensor readings, while the RL agent adjusts the control policy dynamically. The technique is experimentally verified for a piezoelectric-actuated cantilever beam subjected to multimodal disturbances. The hybrid RL-Fuzzy controller achieved a 34.0% reduction in settling time (95% CI: 31.2–36.8%; and the p < 0.001) compared to baseline practices. The hybrid RL-Fuzzy controller lowered the Root-Mean-Square (RMS) acceleration by 28% and was less susceptible to actuator saturation and thermal drift. The proposed framework significantly outperforms traditional PID and LQR controllers and offers a scalable solution to vibration control for smart structures. Its versatility to various systems (e.g., vehicle suspensions, wind turbines) with little retraining demonstrates its potential for practical application.

Keywords—vibration, Reinforcement Learning (RL)_Fuzzy controller, Deep Deterministic Policy Gradient (DDPG), Proportional-Integral-Derivative (PID) controller, piezo

Cite: Taha H. Karam Alsayyad and Baraa M.H. Albaghdadi, "Reinforcement Learning-Based Adaptive Vibration Control for Smart Structures with Fuzzy Uncertainty Quantification," International Journal of Mechanical Engineering and Robotics Research, Vol. 14, No. 6, pp. 608-616, 2025. doi: 10.18178/ijmerr.14.6.608-616

Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).