Title :
The EPSILON processor card: a framework for analog neural computation
Author :
Jackson, Geoff ; Hamilton, Alister ; Murray, Alan F.
Author_Institution :
Dept. of Electr. Eng., Edinburgh Univ., UK
Abstract :
This paper presents new advances in pulse stream neural processing. These advances are made possible by a new neural chip EPSILON II. EPSILON II is made using a standard digital ES2 process and though essentially analog in nature exhibits process invariance. The new chip has been designed specifically with a system level view in mind. The EPSILON Processor Card (EPC) a powerful neural processing tool with the EPSILON II chip at its heart is presented for the first time. Multiple EPC´s can be used in parallel to implement large networks and can take inputs directly from the analog world. To demonstrate the flexibility of the EPC an implementation of an autonomous mobile robot is under construction using the card
Keywords :
neural chips; EPSILON II; EPSILON processor card; analog neural computation framework; autonomous mobile robot; neural chip; pulse stream neural processing; standard digital ES2 process; Analog computers; Circuits; Distributed feedback devices; Heart; Mobile robots; Modular construction; Neural network hardware; Pulse modulation; Very large scale integration; Voltage;
Conference_Titel :
Microelectronics for Neural Networks and Fuzzy Systems, 1994., Proceedings of the Fourth International Conference on
Conference_Location :
Turin
Print_ISBN :
0-8186-6710-9
DOI :
10.1109/ICMNN.1994.593721