Adaptive Neural Network-Based Tracking Control for a Single-Link Flexible Manipulator Under State Constraints.

Area of Science:

• Robotics and Control Systems
• Nonlinear Dynamics
• Fractional Calculus

Background:

• Flexible manipulators offer advantages in lightweight design and energy efficiency for precision tasks.
• Controlling flexible manipulators is challenging due to inherent nonlinearities and system uncertainties.

Purpose of the Study:

• To develop a robust control scheme for trajectory tracking of a single-link flexible manipulator (SLFM) under symmetric time-varying full-state constraints.
• To address the control challenges posed by nonlinearities and uncertainties in SLFM systems.

Main Methods:

• Established a fractional-order dynamic model to capture SLFM characteristics.
• Developed an adaptive radial basis function (RBF) neural network control within a backstepping framework.
• Incorporated a symmetric time-varying barrier Lyapunov function (BLF) and command filters to ensure state constraints and avoid complexity explosion.

Main Results:

• Theoretical analysis confirmed boundedness of all closed-loop system signals and convergence of tracking error.
• Simulations demonstrated excellent control performance with tracking error < 0.02 rad and tip polarization angle < 0.045 rad.
• The proposed controller showed superior performance with less tracking error compared to DSC and SMC methods.

Conclusions:

• The fractional-order adaptive neural network control scheme effectively manages SLFM trajectory tracking under constraints.
• The method validates the effectiveness and superiority of the proposed control strategy for flexible manipulator systems.