Design of controller for nonlinear multivariable system using dynamic neural unit model

The conventional neural network models are a parody of biological neural structures, and have very slow learning speed. In order to emulate some dynamic functions, such as learning and adaptation, and to better reflect the dynamics of biological neurons, Gupta and Rao (1992) have developed a 'dynamic neural unit' (DNU). The proposed neural unit model is to introduce some dynamics to the neural transfer function. Integrating a dynamic elementary processor within the neuron allows the neuron to act dynamic response. Numerical examples are presented for a model system and truck backer-upper control system. These case studies showed that the proposed DNU is very useful in practical applications.