Regresar

Wireless OFDM links with equalizers based on extreme learning machines

Abstract:

Technologies used in wireless access telecommunications networks, particularly those of mobile telephony, are constantly advancing, achieving this through the development of new techniques and methods that allow reaching an adequate level in very important aspects for a communication system. This article presents the extreme learning machine (ELM) with regularized parameter as a suitable alternative to perform channel equalization in orthogonal frequency division multiplexing (OFDM) schemes subject to standard wireless communications. To study their performance, other extreme learning machines proposed as equalizers are also considered. For various signal-to-noise ratios (SNR) and diverse model channels, the bit error rates are exposed, by showing that the superiority of a certain method depends on the adopted wireless link.