Calculation of PID controller parameters by using a fuzzy neural network

Area of Science:

• Control Systems Engineering
• Artificial Intelligence in Engineering
• Automation and Robotics

Background:

• Traditional PID controller tuning often relies on theoretical or numerical methods, which can be complex and time-consuming.
• Achieving optimal performance requires balancing competing criteria like minimizing error and maintaining system stability (sensitivity).

Purpose of the Study:

• To develop an automated formula for designing Proportional-Integral-Derivative (PID) controllers using a Fuzzy Neural Network (FNN).
• To ensure the designed PID controller meets specific performance criteria: minimum Integrated Absolute Error (IAE) and maximum Sensitivity (Ms).

Main Methods:

• Utilized a Fuzzy Neural Network (FNN) to model the relationship between plant parameters and PID controller gains.
• Applied the dominant pole assignment method to simplify the optimization process for tuning rules.
• Developed an FNN-based formula for automatic PID controller tuning, eliminating the need for manual theoretical or numerical approaches.

Main Results:

• The FNN system successfully identified the plant model and controller parameter relationships based on IAE and Ms.
• The developed FNN-based formula enables automatic tuning of PID controllers for varying system parameters.
• The approach demonstrated effectiveness in a motor position control simulation, showing adaptability to system model changes.

Conclusions:

• The proposed FNN-based approach provides an effective and automated method for designing PID controllers.
• This technique simplifies the tuning process while optimizing for key performance metrics like IAE and Ms.
• The FNN-based formula offers a robust solution for adaptive controller modification in dynamic systems.