NESP: an analog NEural Signal Processor

This paper presents a signal processor based on neural computation rather than traditional programming. This processor has analog input-output, analog processing unit and analog memory. It doesn't require any converter to be interfaced to analog signals. Configuration data is stored into an internal analog memory, based on flash EEPROM technology. A modular architecture allows one to put together more processors to configure more layers and more neurons.

[1]  Eric A. Vittoz,et al.  Analog VLSI implementation of neural networks , 1990, IEEE International Symposium on Circuits and Systems.

[2]  Michiel Steyaert,et al.  A high-dynamic-range CMOS op amp with low-distortion output structure , 1987 .

[3]  Z. Weinberg,et al.  On tunneling in metal‐oxide‐silicon structures , 1982 .

[4]  B. Eitan,et al.  Analysis and modeling of floating-gate EEPROM cells , 1986, IEEE Transactions on Electron Devices.

[5]  M. Lenzlinger,et al.  Fowler‐Nordheim Tunneling into Thermally Grown SiO2 , 1969 .

[6]  V. Hu,et al.  EEPROM device as a reconfigurable analog element for neural networks , 1989, International Technical Digest on Electron Devices Meeting.

[7]  S. Y. Kung,et al.  Digital VLSI architectures for neural networks , 1989, IEEE International Symposium on Circuits and Systems,.