Abstract
Kohonen (1988) introduced a new type of neural network that exhibits very interesting pattern classification and clustering properties. So far, they have been successfully used to solve various problems in fields like image processing or robot control. Unfortunately, although the task each cell has to accomplish is rather simple, computing time of software simulations becomes prohibitive as the size of a network increases. For real-time applications, parallel processing is required. Analog circuits are believed to allow potentially highest density of integration, and highest speed. However, since density is essential, only poor accuracy may be achieved. It is therefore necessary to select architectures that guarantee proper behaviour of the network despite of all error sources that may affect it.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Heim, P., Hochet, B. and Vittoz, E., “Generation of Learning Neighbourhood in Kohonen Feature Map by Means of Simple Nonlinear Network”, Electronics Letters, January 1991, Vol. 27, No 3, pp. 275–277
Heim, P., private communication, 1991
Kohonen, T., Self-Organisation and Associative Memory, Springer Verlag 1988, ch. 5, pp. 119–157
Lazzaro, J., Ryckebush, S., Mahowald, M. A. and Mead, C., “Winner-Take-All networks of order N complexity”, Proc. 1988 IEEE conference on neural information processing — natural and synthetic, Denver, 1988, pp. 703–711
Landolt, O., Vittoz, E. and Heim, P., “CMOS self-biased Euclidean distance computing circuit with high dynamic range”, Electronics Letters, February 1992, Vol. 28, No 4, pp. 352–354
Säckinger, E. and Guggenbühl, W., “An analog trimming circuit based on a floating-gate device”, IEEE Journal of Solid State Circuits, Vol. SC-23, No. 6, December 1988, pp. 1437–1440
Sorouchyari, E., “Fractal dimension computation based on a self-organizing neural network”, to be published
Vittoz, E. et al, “Analog storage of adjustable synaptic weights”, Proc. ITG/IEEE workshop on microelectronics for neural networks, Dortmund, Germany, 1990
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1994 Springer Science+Business Media New York
About this chapter
Cite this chapter
Landolt, O. (1994). An Analog CMOS Implementation of a Kohonen Network with Learning Capability. In: Delgado-Frias, J.G., Moore, W.R. (eds) VLSI for Neural Networks and Artificial Intelligence. Springer, Boston, MA. https://doi.org/10.1007/978-1-4899-1331-9_2
Download citation
DOI: https://doi.org/10.1007/978-1-4899-1331-9_2
Publisher Name: Springer, Boston, MA
Print ISBN: 978-1-4899-1333-3
Online ISBN: 978-1-4899-1331-9
eBook Packages: Springer Book Archive