Design of twin delayed deep deterministic policy gradient RL based adaptive controller for DC motor speed regulation...

Area of Science:

• Control Systems Engineering
• Machine Learning
• Robotics

Background:

• Reinforcement learning (RL) excels at optimizing complex decision-making.
• DC motor speed control is crucial in many industrial applications.
• Environmental uncertainties challenge traditional control methods.

Purpose of the Study:

• To develop and evaluate a TD3 RL-based adaptive speed controller for DC motors.
• To compare the TD3 controller's performance against benchmark controllers.
• To assess controller robustness and adaptability under dynamic environmental conditions.

Main Methods:

• Implementation of a Twin Delayed Deep Deterministic (TD3) policy gradient RL algorithm.
• Development of an adaptive speed controller for a DC motor model.
• Comparative analysis using benchmark controller techniques.
• Performance evaluation through simulations with constant and variable speed setpoints.

Main Results:

• The TD3 controller exhibited enhanced adaptability and robustness.
• Quantitative analysis using error indices (ISE, ITAE, IAE, ITSE) confirmed superior performance.
• The controller effectively minimized errors in dynamic operating conditions.

Conclusions:

• TD3 RL provides an efficient and adaptive solution for DC motor speed control.
• The proposed controller demonstrates significant advantages over traditional methods in uncertain environments.
• This approach offers precise speed regulation and error minimization for DC motors.