Title :
An Experimental Study on Nonlinear Function Computation for Neural/Fuzzy Hardware Design
Author :
Basterretxea, Koldo ; Tarela, José Manuel ; Del Campo, Inés ; Bosque, Guillermo
Author_Institution :
Dept. of Electron. & Telecommun., Univ. of the Basque Country, Bilbao
Abstract :
An experimental study on the influence of the computation of basic nodal nonlinear functions on the performance of (NFSs) is described in this paper. Systems´ architecture size, their approximation capability, and the smoothness of provided mappings are used as performance indexes for this comparative paper. Two widely used kernel functions, the sigmoid-logistic function and the Gaussian function, are analyzed by their computation through an accuracy-controllable approximation algorithm designed for hardware implementation. Two artificial neural network (ANN) paradigms are selected for the analysis: backpropagation neural networks (BPNNs) with one hidden layer and radial basis function (RBF) networks. Extensive simulation of simple benchmark approximation problems is used in order to achieve generalizable conclusions. For the performance analysis of fuzzy systems, a functional equivalence theorem is used to extend obtained results to fuzzy inference systems (FISs). Finally, the adaptive neurofuzzy inference system (ANFIS) paradigm is used to observe the behavior of neurofuzzy systems with learning capabilities
Keywords :
Gaussian processes; adaptive systems; backpropagation; fuzzy reasoning; nonlinear functions; radial basis function networks; Gaussian function; adaptive neurofuzzy inference system; artificial neural network; backpropagation neural networks; functional equivalence theorem; neural-fuzzy hardware design; nonlinear function computation; performance index; radial basis function networks; sigmoid-logistic function; Algorithm design and analysis; Approximation algorithms; Artificial neural networks; Backpropagation algorithms; Computational modeling; Computer architecture; Fuzzy systems; Hardware; Kernel; Performance analysis; Approximation capability; Gaussian function; centered recursive interpolation (CRI); neurofuzzy hardware; sigmoid function; Computer-Aided Design; Electronics; Equipment Design; Equipment Failure Analysis; Fuzzy Logic; Neural Networks (Computer); Nonlinear Dynamics; Signal Processing, Computer-Assisted;
Journal_Title :
Neural Networks, IEEE Transactions on
DOI :
10.1109/TNN.2006.884680
