Adaptive Neural PID Controllers in Modeling and Control of Collaborative Robots: Analysis, Comparison and Application Recommendations

Abstract

The work investigates adaptive Neural PID controllers as an intelligent control method for collaborative robots operating in Industry 5.0 cyber-physical environments. The study analyzes classical PID controllers, adaptive PID controllers, neural network-based PID controllers, and hybrid Neuro-PID control systems. Their mathematical models, stability properties, adaptation capability, computational complexity, and suitability for Digital Twin systems are compared. Simulation-based evaluation demonstrates that Neural PID controllers significantly improve control accuracy, robustness, and adaptability in dynamic environments. Practical recommendations are provided for their implementation in collaborative robotic systems.