A Vibration Control Method for Hybrid-Structured Flexible Manipulator Based on Sliding Mode Control and Reinforcement...

Area of Science:

• Robotics
• Control Systems Engineering

Background:

• Hybrid-structured flexible manipulators exhibit complex dynamics and strong coupling, making vibration suppression challenging.
• Natural frequencies in these systems vary with telescopic joint motion, complicating real-time control.
• Existing methods struggle with rapid and accurate vibration suppression in such complex robotic systems.

Purpose of the Study:

• To develop a combined control strategy for hybrid-structured flexible manipulators.
• To enhance tip positioning accuracy and trajectory tracking performance.
• To effectively suppress vibrations in flexible robotic systems.

Main Methods:

• Decomposition of the tip state signal into elastic vibration and equilibrium position signals.
• Implementation of an improved nominal model-based sliding mode controller (NMBSMC) as the main controller.
• Integration of an actor-critic-based reinforcement learning controller (ACBRLC) as an auxiliary controller for compensation torque.
• Utilizing prioritized experience replay within the ACBRLC for efficient learning.

Main Results:

• The improved NMBSMC effectively tracks the vibration equilibrium position by combining nominal and practical model-based approaches.
• The ACBRLC generates precise compensation torque using elastic vibration signals to counteract residual vibrations.
• Experimental results demonstrate significant improvements in tip positioning and trajectory tracking accuracy.
• The combined control method shows robustness in suppressing vibrations for flexible manipulators.

Conclusions:

• The proposed combined control method effectively addresses the challenges of vibration suppression in hybrid-structured flexible manipulators.
• The synergistic action of NMBSMC and ACBRLC leads to superior positioning and tracking accuracy.
• This approach offers a robust solution for enhancing the performance of flexible robotic systems.