Optimizing the Tuning of Fuzzy-PID Controllers for Motion Control of Friction Stir Welding Robots

Friction stir welding (FSW) is defined as a solid-state welding method that is required to be accurate, especially for its motion. This requirement can be satisfied by implementing an accurate controller. The aim of this research was to develop an accurate control system based on a fuzzy-proportional integral derivative (PID) controller for parallel manipulator FSW robots. In order to achieve a higher accuracy in motion control, the tuning optimisation process for a fuzzy-PID controller was conducted using a genetic algorithm (GA) and particle swarm optimisation (PSO). The optimisation algorithms were applied to simultaneously tune the fuzzy rules and output of the membership function from the fuzzy inference system (FIS). The PID controller was designed and tuned using a MATLAB® PID Tuner to obtain the desired response. It was then developed into a fuzzy-PID controller with Sugeno type-1 FIS with 2 inputs and 1 output. The tuning optimisation of the fuzzy-PID controller using GA and PSO was performed to achieve the global minimum integral absolute error (IAE) of the angular velocity. MATLAB® Simulink® was employed to test and simulate the controllers for three motors in the FSW robot model. The IAE values of the PID controller implemented for each motor were 0.03644, 0.04893, and 0.04893. The IAEs of the implemented fuzzy-PID-GA (output and rules) controller were 2.061, 2.048, and 2.048; of the implemented fuzzy-PID-GA (output) controller were 0.03768, 0.05059, and 0.05059; of the fuzzy-PID-PSO (output and rules) controller were 0.01886, 0.0253, and 0.02533; and of the fuzzy-PID-PSO (output) controller were 0.03767, 0.05059, and 0.05059. Therefore, the fuzzy-PID-PSO (output and rules) controller gave the most accurate results and outperformed the others.