Tensor based receivers for nonlinear Radio Over Fiber uplinks in multiuser CDMA systems

Abstract

International audienceTwo semi-blind tensor based receivers for uplink radio over fiber (ROF) nonlinear instantaneous channels in a multiuser code division multiple access (CDMA) environment are proposed in this paper. These receivers are based on the PARAFAC decomposition of a tensor (three-way array) composed of received signals with space, time and code diversities. A great advantage of the tensor approach is that it allows joint blind estimation of the channel and transmitted signals under weak uniqueness conditions compared with existing methods, which is particularly suitable when identifying nonlinear systems. Two estimation algorithms are proposed and illustrated by means of computer simulations

Processamento Tensorial de Sinais e Geometria da Informação: Novos Caminhos para os Sistemas de Comunicações Digitais

“…Técnicas de estimação de canal e cancelamento de interferências em canais MIMO não-lineares foram desenvolvidas nosúltimos anos, tendo como base a utilização de tensores construídos a partir dos sinais recebidos [63], [64], [65], bem como a partir de estatísticas de egunda ordem destes sinais [63], [66], [64], [67]. Estas abordagens provêem condições de unicidade mais suaves que as do caso matricial, permitindo um tratamento com um número reduzido de antenas receptoras.…”

MIMO Volterra Modeling for Nonlinear Communication Channels

“…Discrete-time MIMO Volterra models have being studied in a few works. However, some of these works deal with MIMO Volterra systems less generical than (53) [8,30,36,38,39,41,43], while some others deal with a real-valued version [42] or with a memoryless version of this model [57][58][59][60][61][62]. To the best of our knowledge, the techniques that we proposed in [63,64] are only works that use the generical complex-valued equivalent baseband MIMO Volterra model (53).…”