Title :
Analog decoding using a gradient-type neural network
Author :
Ciocoiu, Iulian B.
Author_Institution :
Fac. of Electron. & Telecommun., Tech. Univ. Iasi, Romania
fDate :
7/1/1996 12:00:00 AM
Abstract :
The problem of analog (soft) decision decoding of block codes by means of neural networks is addressed. The proposed solution is based on a recurrent high-order network implementing a special gradient-type system. Simulation results for two different codes are reported, showing improved performances over the classical hard decision decoder
Keywords :
block codes; decoding; recurrent neural nets; analog decision decoding; block codes; gradient-type neural network; recurrent high-order network; Artificial neural networks; Block codes; Circuits; Decoding; Frequency; Hamming distance; Neural networks; Neurons; Optical computing; Optical signal processing;
Journal_Title :
Neural Networks, IEEE Transactions on
