Compound FAT-based prespecified performance learning control of robotic manipulators with actuator dynamics

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence in Engineering
• Applied Mathematics

Background:

• Electrically-driven robotic manipulators often face challenges with unknown system dynamics and output constraints.
• Existing control methods may struggle to simultaneously address approximation accuracy and strict output limitations.
• Backstepping algorithms provide a foundation but require enhancements for complex robotic systems.

Purpose of the Study:

• To develop a novel function approximation technique (FAT)-based compound learning control law for electrically-driven robotic manipulators.
• To address the challenge of output constraints in robotic manipulator control.
• To ensure system stability and bounded tracking errors through advanced control strategies.

Main Methods:

• Utilizing a backstepping algorithm framework combined with a new function approximation technique (FAT).
• Employing Fourier series expansion for approximating unknown system terms.
• Integrating error transformation, performance functions, and dynamic surface control to handle output constraints.
• Developing novel compound adaptation laws.

Main Results:

• The proposed compound learning controller guarantees that all system signals are uniformly ultimately bounded.
• Tracking errors are maintained within predefined bounds during robotic manipulator operation.
• The function approximation technique (FAT) demonstrates accuracy in identifying unknown system dynamics.

Conclusions:

• The developed control law effectively manages output constraints and unknown dynamics in robotic manipulators.
• The approach ensures robust and stable performance, confirmed through simulations.
• This method offers a significant advancement in the control of electrically-driven robotic systems.