Remarks on Hypercomplex-Valued Neural Network-Based Direct Adaptive Controller

In this study, we investigate the control performance of an adaptive controller using multilayer hypercomplex-valued neural networks, such as complex, hyperbolic, bicomplex and quaternion neural networks. The control system is a direct controller whereby the control input of a plant is synthesized online using the multilayer hypercomplex-valued neural network to track the plant output to the desired output generated by a reference model. Computational experiments to control a multiple-input and multiple-output discrete-time nonlinear plant are conducted to evaluate the capability and characteristics of the hypercomplex-valued neural network-based direct adaptive controller. Experimental results show the feasibility and effectiveness of the proposed controller.