A Hybrid Algorithm for Adaptive Neuro-controllers

Abstract

In this study, a novel hybrid algorithm consisting of the least mean square and backpropagation neural network is proposed to auto-adjust adaptive proportional integral derivative (PID) controller gains for improving the transient response of linear systems. The hybrid approach comprises the scheme of the two algorithms running in parallel and updates PID gains simultaneously. All algorithms are implemented on the same linear system and present a general framework for different scenarios such as initial PID gains, learning rates, and target functions. The results show that the presented hybrid algorithm has better accuracy, precision, F1-score, adaptability, and robustness than origin algorithms, and significantly improves the controllability in most of the system scenarios. It also exhibits better performance in periodic incremental and decremental targets compared to origin algorithms. Different hybridization levels are also simulated and are highlighted as significant features of their performance. This work can be expanded to the combination of other well-known algorithms, paving the way to significant improvements in control system applications.