Neural Network-Based Integral Sliding Mode Boundary Control for Flexible Manipulators with Boundary Perturbations

This paper presents a boundary sliding mode control strategy integrated with neural networks to address boundary disturbances in a flexible arm system. A sliding surface and corresponding boundary controller are designed based on boundary measurements. Neural networks are employed to approximate unknown disturbances, while a hyperbolic tangent function is introduced to mitigate control chattering. Lyapunov stability analysis demonstrates that the systems energy ensures finite-time convergence to the sliding surface, followed by exponential convergence of the system states along the surface. Numerical simulations are conducted to validate the effectiveness of the proposed boundary sliding mode control approach.