Simple Neural Network Compact Form Model-Free Adaptive Controller for Thin McKibben Muscle System

This paper proposes a simple neural network compact form model-free adaptive controller (NNCFMFAC) for single thin McKibben muscle (TMM) system. The main contribution of this work is the simplification current (NN) based (CFMFAC), which requires only two weights. achieved by designing NN topology specifically to enhance CFMFAC response. prominent control parameters are combined and one weight used self-tuning, while second minimize offset at each operating point. Hence issues redundant weights in complex neuro CFMFACs slow response significantly addressed. idea proven three ways: analytically, simulation on nonlinear system experiments TMM platform. Experimental results demonstrate superiority proposed method over conventional confirmed 76% improvement convergence speed 60% reduction root mean square error (RMSE). It envisaged that can be very useful driven applications as it model independent, has fast response, high tracking accuracy, minimal complexity.