Skip to Main content Skip to Navigation

Nonlinear MIMO communication systems : channel estimation and information recovery using Volterra models

Abstract : Due to the presence of nonlinear devices such as power amplifiers (PA) and optical instruments, the communication signals are sometimes corrupted by nonlinear distortions. In such cases, nonlinear models are used to provide an accurate signal representation, allowing the development of efficient signal processing techniques capable of eliminating or reducing these nonlinear distortions. In this context, the choice of the nonlinear system model plays a fundamental role. The Volterra model has since longtime been used to represent communication systems in presence of nonlinear distortions, with applications for modeling satellite communication links, orthogonal frequency division multiplexing (OFDM) systems and radio over fiber (ROF) channels. The main objective of this thesis is to propose techniques for channel estimation and information recovery in multiple-input-multiple-output (MIMO) Volterra communication systems. This kind of MIMO model is able of modeling nonlinear communication channels with multiple transmit and receive antennas, as well as multi-user channels with a single transmit antenna for each user and multiple receive antennas. Channel estimation and equalization techniques are developed for three types of nonlinear MIMO communication systems: OFDM, ROF and Code division multiple access (CDMA)-ROF systems. According to the considered communication systems, different kinds of MIMO Volterra models are used. In the case of OFDM systems, we develop receivers that exploit the diversity provided by a proposed transmission scheme. In the case of time and space division multiple access (TDMA-SDMA) systems, a set of orthonormal polynomials is developed for increasing the convergence speed of a supervised adaptive MIMO Volterra estimation algorithm. Moreover, in order to develop signal processing techniques for MIMO Volterra communication channels in a blind scenario, we make use of tensor decompositions. By exploiting the fact that Volterra models are linear with respect to their coefficients, blind estimation and equalization of MIMO Volterra channels are carried out by means of the Parallel Factor (PARAFAC) tensor decomposition, considering TDMA-SDMA and CDMA communication systems.
Complete list of metadatas
Contributor : Estelle Nivault <>
Submitted on : Friday, February 26, 2010 - 1:47:01 PM
Last modification on : Wednesday, October 14, 2020 - 4:23:30 AM
Long-term archiving on: : Thursday, October 18, 2012 - 4:10:46 PM


Files produced by the author(s)


  • HAL Id : tel-00460160, version 1



Carlos Alexandre Rolim Fernandes. Nonlinear MIMO communication systems : channel estimation and information recovery using Volterra models. Networking and Internet Architecture [cs.NI]. Université de Nice Sophia Antipolis, 2009. English. ⟨tel-00460160⟩



Record views


Files downloads