An analogue circuit implementation is presented for an adaptive resonance theory neural network architecture, called the augmented ART-I neural network (AARTI-NN). The AARTI-NN is a modification of the popular ARTI-NN, developed by Carpenter and Grossberg, and it exhibits the same behaviour as the ART1-NN. The AARTI-NN is a real-time model, and has the ability to classify an arbitrary set of binary input patterns into different clusters. The design of the AARTI-NN circuit is based on a set of coupled nonlinear differential equations that constitute the AARTI-NN model. The circuit is implemented by utilizing analogue electronic components such as operational amplifiers, transistors, capacitors, and resistors. The implemented circuit is verified using the PSpice circuit simulator, running on Sun workstations. Results obtained from the PSpice circuit simulation compare favourably with simulation results produced by solving the differential equations numerically. The prototype system developed here can be used as a building block for larger AARTI-NN architectures, as well as for other types of ART architectures that involve the AARTI-NN model.
IntroductionInformation processing and management in modern military and commercial systems is growing more complex and is requiring higher performance. Traditional computational approaches have not met the challenge posed by these expanding requirements. Neural networks offer new computational solutions to these requirements that are potentially robust to noise and incomplete information, adaptive to a changing environment, intrinsically massively parallel and fault tolerant. Neural networks are currently realized in a number of ways. Most of the neural networks in the open literature have been implemented via computer simulation of their corresponding mathematical models. For real-life applications, however, neural networks need to be implemented as analogue, digital, or hybrid (analogue digital) hardware.In this paper we focus our attention on the hardware implementation of an adaptive resonance theory neural network named AARTI-NN (augmented adaptive resonance theory-l neural network). The AARTI-NN was developed by Heileman er al. (1992), and is a modification of the popular ARTI-NN introduced by Carpenter and Grossberg (1987). The major difference between the AARTI-NN and the ARTI-NN is that the AARTI-NN is completely described by a set of differential equations, while the ARTI-NN incorporates algorithmic components in its description. As Carpenter (1989) points out, differential equations constitute the language of real-time models. Hence, from this perspective the AARTI-NN is a real-time
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.