Please use this identifier to cite or link to this item:
|Title:||Theoretical analysis of nonlinear amplification effects in massive MIMO systems|
Barradas, Filipe M.
Cabral, Pedro M.
|Keywords:||Hybrid analogic/digital architectures|
|Abstract:||To fulfill 5th Generation (5G) communication capacity demands, the use of a large number of antennas has been widely investigated, and the array gain and spatial multiplexing that are offered by massive multiple input multiple output (mMIMO) have been used to improve the capacity. Fully digital architectures are not feasible for a large number of antennas, and hybrid analog/digital systems have emerged as options to retain a high number of antennas without as many radio frequency (RF) chains. However, these systems have, as consequences, non-avoidable nonlinear effects due to power amplifiers functioning in nonlinear regions. The strong nonlinear effects throughout the transmission chain will have a negative impact on the overall system’s performance. Being able to access this impact is very important. For this purpose, we propose analytical and semi-analytical tools that allow for the evaluation of the nonlinear effects of a hybrid analog/digital orthogonal frequency-division multiplexing (OFDM) system. The proposed analysis starts with the characterization of the power amplifier’s (PA) nonlinear response. This response is then used to derive a semi-analytic bit error rate expression. The theoretical tools are validated by using numerical results from two different cases: in the first one, the nonlinear PA response is assumed to follow an analytical model found in the literature and, in the second, the used nonlinear polynomial model mimics the response of a real amplifier. Using these two scenarios, the proposed tools are shown to be accurate making it possible to predict the nonlinearities’ penalties in hybrid analog/digital OFDM systems and/or to assess the optimal operation point for a specific nonlinear amplifier.|
|Appears in Collections:||DETI - Artigos|
IT - Artigos
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.