Neural network-based adaptive second-order sliding mode control for uncertain manipulator systems with input...

Area of Science:

• Robotics
• Control Systems Engineering
• Artificial Intelligence

Background:

• Robotic manipulators face challenges like dynamic uncertainty, external disturbances, and input saturation.
• Existing sliding mode control methods can suffer from chattering and may not handle unknown nonlinearities effectively.

Purpose of the Study:

• To propose a novel control scheme for robotic manipulators addressing trajectory tracking problems.
• To enhance robustness, convergence speed, and tracking accuracy under uncertain and disturbed conditions.

Main Methods:

• Design of a model-based second-order non-singular fast terminal sliding mode controller (SONFTSMC) to mitigate chattering.
• Development of a fuzzy wavelet neural network (FWNN) for adaptive estimation of lumped unknown nonlinear uncertainties.
• Autonomous parameter adjustment of the FWNN using an adaptive method.

Main Results:

• The proposed second-order non-singular fast terminal sliding mode (SONFTSM) control improves convergence speed and tracking accuracy.
• The control strategy demonstrates enhanced robustness against dynamic uncertainty and external disturbances.
• Comparative simulations validate the advantages of the SONFTSM strategy over existing methods.

Conclusions:

• The novel SONFTSM control scheme effectively solves trajectory tracking problems for robotic manipulators.
• The integration of FWNN provides robust estimation of uncertainties, enhancing overall system performance.
• This approach offers a significant advancement in sliding mode control for complex robotic systems.