Previous Issues

*CVSSP, School of EEIT&M, University of Surrey, Guildford, Surrey GU2 5XH, U.K

** Electrical & Computer Engineering Department, Kuwait University, P0 Box 5969, Safat 13060, Kuwait. E-mail: sfoda@eng.kuniv.edu.kw 

ABSTRACT

This paper describes an artificial neural network model that can detect and correct linear Hamming codes. An error-correcting neural network decoder is designed using two single layer networks. The first layer is a linear Perception while the other layer is an Outstar vector generator. Other neural network models that can be used to decode code words with the ability to correct single bit errors are simulated and compared. The proposed ANN exhibits a simple structure and efficient performance which makes it a suitable candidate for real-time implementation.