A Study on a Variable-Gain PID Control for a Pneumatic Servo System Using an Optimized PSO-Type Neural Network

• Published in: Actuators Vol. 14; no. 5; p. 250
• Main Authors: Mu, Shenglin, Shibata, Satoru, Baba, Daisuke, Oshita, Rikuto
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.05.2025
• Subjects: Accuracy; Algorithms; Control systems; Design; Genetic algorithms; Mathematical optimization; Methods; Neural networks; Nonlinearity; Optimization; optimized particle swarm optimization; Particle swarm optimization; PID control; pneumatic servo system; Pneumatics; Proportional integral derivative; Robust control; Servocontrol; Simulation; Systems stability; Variable gain
• ISSN: 2076-0825; 2076-0825
• DOI: 10.3390/act14050250

This study investigates the application of proportional–integral–derivative (PID) control enhanced with an optimized particle swarm optimization (OPSO)-type neural network (NN) to improve the control performance of a pneumatic servo system. Traditional PID methods struggle with inherent nonlinearities in pneumatic servo systems. To address this limitation, we integrate two OPSO-type NNs within the PID framework, thereby developing a robust control strategy that compensates for these nonlinear characteristics. The OPSO-type NNs are particularly efficient in solving complex optimization problems without requiring differential information, demonstrating superior simplicity and efficacy compared to traditional methods, such as genetic algorithms. In our approach, one of the OPSO-type NNs is utilized to tune the PID controller gains, while the other adjusts the control output. The experimental results show that the proposed method enhances the position control accuracy of the pneumatic servo system. Furthermore, this approach holds promise for improving the responsiveness, stability, and disturbance suppression capabilities of pneumatic systems, paving the way for advanced control applications in this field.