An Experimental Study on Nonlinear Function Computation for Neural/Fuzzy Hardware Design

• Published in: IEEE transaction on neural networks and learning systems 2007-01, Vol.18 (1), p.266-283
• Main Authors: Basterretxea, K., Tarela, J.M., del Campo, I., Bosque, G.
• Format: Article
• Language: English
• Subjects: Algorithm design and analysis; Algorithms; Applied sciences; Approximation algorithms; Approximation capability; Artificial intelligence; Artificial neural networks; Backpropagation algorithms; centered recursive interpolation (CRI); Computational modeling; Computer architecture; Computer science; control theory; systems; Computer-Aided Design; Connectionism. Neural networks; Electronics; Equipment Design; Equipment Failure Analysis; Exact sciences and technology; Fuzzy Logic; Fuzzy systems; Gaussian function; Hardware; Kernel; Neural networks; Neural Networks (Computer); neurofuzzy hardware; Nonlinear Dynamics; Performance analysis; sigmoid function; Signal Processing, Computer-Assisted - instrumentation; Studies
• ISSN: 1045-9227; 2162-237X; 1941-0093; 2162-2388
• DOI: 10.1109/TNN.2006.884680

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