Comparison of analog and digital nanosystems: Issues for the nano-architect

Improvements in manufacturing and assembly of novel devices such as the carbon nanotube (CNT) and semiconductor nanowires has led to the exploration of new nano-architectures utilizing such devices for applications such as general purpose computing, image processing etc. This paper discusses different possible nanoscale implementations of cellular neural networks (CNN). It is seen that while an analog nanoscale implementation of the CNN may be difficult with self-assembly based approaches given the requirements for customization of devices and arbitrary routing, a digital equivalent may be realizable in the near term. One such digital CNN design based on the NASIC nanoscale fabric is shown. A specialized architecture for CNN with resonant tunneling diodes (RTD) is also discussed.