A neural analog-to-digital converter with resolution enhancement

Abstract

A neural-based analog-todigital (AD) converter was first proposed by Tank and Hopfield in 1986. Hopfield applications have typically used high-gain amplifiers for the neurom. Unfortunately, this led to missing output codes and unequal step widths in the A D converter application. n e new approach proposed here use lowgain analog amplifiers for the arti!cial neurons. This allows additional bits of information to be extracted from the A D converter output. In this paper, we discuss the hardware implementation of our resolution enhancement method applied to a four-bit neural A D converter. R e method yielded an increase in resolution from 4 to 8 bits.