Study on single input-single output of CMAC neural network

Considering of the real question in the conventional Cerebllar Model Articulation Controller (CMAC) neural network, a single input-single output CMAC (SISO CMAC) algorithm is studied in this paper. The direct weight address mapping techniques are used in the algorithm, and the relation between the inputs and the weights is established by taking the quantification of input x as their weight address, and the generalization capability of network is enhanced by input x normalized. The study of network uses random sample, so the network has characteristics of simple study and quick convergence and high accuracy of function approximation, etc, and implement in single chip microcomputer. In the end, it is showed the algorithm is validity by theory proved and simulation experiments.