Coot optimization algorithm-tuned neural network-enhanced PID controllers for robust trajectory tracking of...

Area of Science:

• Robotics
• Control Systems Engineering
• Artificial Intelligence

Background:

• Robotic manipulators are complex, nonlinear systems susceptible to disturbances and uncertainties.
• Effective control is crucial for achieving precise trajectory tracking and stable operation.
• Existing control methods often struggle with the inherent complexities of robotic manipulators.

Purpose of the Study:

• To develop and evaluate advanced control strategies for a three-Link Rigid Robot Manipulator (3-LRRM).
• To address the trajectory tracking problem by integrating Neural Network (NN) with Proportional, Integral, and Derivative (PID) control.
• To optimize controller parameters using the Coot Optimization Algorithm (COOA) for improved performance and reduced signal chattering.

Main Methods:

• Design of three distinct control structures: Neural Controller Like PIPD (NN-PIPD), Neural Network plus PID (NN+PID), and Elman Neural Network Like PID (ELNN-PID).
• Parameter tuning of controllers using the Coot Optimization Algorithm (COOA) to minimize Integral Time Square Error (ITSE).
• Introduction of a novel objective function to minimize control signal chattering during the tuning process.

Main Results:

• The NN-PIPD controller exhibited superior trajectory tracking performance compared to NN+PID and ELNN-PID controllers.
• Evaluations demonstrated the controllers' effectiveness in disturbance rejection, handling model uncertainties, and adapting to varying initial conditions.
• The NN-PIPD controller achieved the minimum ITSE of 0.001777, indicating optimal performance.

Conclusions:

• The proposed NN-PIPD controller is highly effective for robotic manipulator trajectory tracking, disturbance rejection, and parameter variation.
• The integration of Neural Networks with PID control, optimized by COOA, offers a robust solution for complex robotic systems.
• The study highlights the potential of NN-PIPD for enhancing the stability and robustness of robotic manipulator control.