Carlos Estêvão Rolim Fernandes

Blind channel identification algorithms based on the Parafac decomposition of cumulant tensors: The single and multiuser cases

In this paper, we exploit the symmetry properties of 4th-order cumulants to develop new blind channel identification algorithms that utilize the parallel factor (Parafac) decomposition of cumulant tensors by solving a single-step (SS) least squares (LS) problem. We first consider the case of single-input single-output (SISO) finite impulse response (FIR) channels and then we extend the results to multiple-input multiple-output (MIMO) instantaneous mixtures. Our approach is based on 4th-order output cumulants only and it is shown to hold for certain underdetermined mixtures, i.e. systems with more sources than sensors. A simplified approach using a reduced-order tensor is also discussed. Computer simulations are provided to assess the performance of the proposed algorithms in both SISO and MIMO cases, comparing them to other existing solutions. Initialization and convergence issues are also addressed.

Blind identification of MISO-FIR channels

In this paper, we address the problem of determining the order of MISO channels by means of a series of hypothesis tests based on scalar statistics. Using estimated 4th-order output cumulants, we exploit the sensitiveness of a Chi-square test statistic to the non-Gaussianity of a stochastic process. This property enables us to detect the order of a linear finite impulse response (FIR) channel. Our approach leads to a new channel order detection method and we provide a performance analysis along with a criterion to establish a decision threshold, according to a desired level of tolerance to false alarm. Afterwards, we introduce the concept of MISO channel nested detectors based on a deflation-type procedure using the 4th-order output cumulants. The nested detector is combined with an estimation algorithm to select the order and estimate the parameters associated with different transmitters composing the MISO channel. By treating successively shorter and shorter channels, it is also possible to determine the number of sources.

Tensor-Based Blind Channel Identification

We propose a blind FIR channel identification method based on the parallel factor (Parafac) analysis of a 3rd-order tensor composed of the 4-th order output cumulants. Our algorithm is based on a single-step least squares (LS) minimization procedure instead of using classical three-step alternating least squares (ALS) methods. Using a Parafac-based decomposition, we avoid any kind of pre-processing such as the prewhitening operation, which is mandatory in most methods using higher-order statistics. Our method retrieves the channel vector without any permutation or scaling ambiguities. In addition, we establish a link between the cumulant tensor decomposition and the joint-diagonalization approach. Computer simulations illustrate the performance gains that our method provides with respect to other classical solutions. Initialization and convergence issues are also addressed.

Blind Multipath MIMO Channel Parameter Estimation Using the Parafac Decomposition

In this paper, we consider the problem of estimating the physical parameters that describe a multipath MIMO communication channel model characterized by specular reflections due to remote scatterers. Using the impulse response channel coefficients, we introduce a channel model based on a 3rd- order tensor structure that admits a Parafac decomposition with rank equal to the number of propagation paths. The alternating least squares (ALS) algorithm is used to estimate the channel spatial and temporal signatures and the multipath parameters are then extracted by means of MUSIC-like subspace algorithms, enabling us to recover the transmit and receive angles as well as the path propagation delays of the MIMO channel, without ambiguities. Computer simulation results are shown to illustrate the performance of the proposed ALS-MUSIC estimation algorithm.

Space-time processing with a decoupled delayed decision-feedback sequence estimator

This work aims at investigating the performance of a decoupled space-time processing structure based on a delayed decision-feedback sequence estimator (D-ST-DDFSE), in the context of the Enhanced Data rates for GSM Evolution (EDGE) system. The main idea in using decoupled space-time (D-ST) processing structures is to separate co-channel interference (CCI) reduction from intersymbol interference (ISI) suppression. Consequently, all degrees of freedom of a space-time (ST) front-end are dedicated to treat CCI, leaving ISI to be suppressed by a temporal equalizer. Due to the 8-PSK modulation and the large delay spread values compared to the symbol period, optimum detection becomes too complex in the EDGE system, which makes DDFSE a promising scheme for ISI suppression. The performance of D-ST-DDFSE is analyzed through link-level simulations under the context of COST 259 channel models for typical urban (TU) and bad urban (BU) propagation scenarios. Improved performance of this D-ST technique over a space-time linear equalizer is observed.

Performance evaluation of sub-space techniques for array processing in TDMA systems

In studying the possibility of increasing wireless system capacity, we evaluate the performance of some algorithms that make use of sub-space techniques to estimate the covariance matrix, and compare these results to those obtained through traditional methods, such as the direct matrix inversion-maximum signal-to-noise ratio and maximal ratio combining. Illustrative simulation results demonstrate that the minimum mean square error-signal sub-space and the weighted sub-space algorithms may lead to a better performance than full-rank conventional algorithms. Furthermore, a more elaborate system-level simulation in a TDMA IS-136 context is performed and it shows that such benefits also appear in a more practical scenario.

BLAST/MIMO performance with space-time processing receivers

The use of antenna arrays at both ends of the link has attracted significant attention of the researches on space-time equalization and coding techniques for so called multiple-input-multiple-output (MIMO) channels. The presence of strong co-channel interference (CCI) in addition to inter-symbol interference (ISI) in current wireless (mobile) communication systems places a significant challenge to MIMO space-time equalizers. We assess the performance of BLAST/MIMO on frequency-selective MIMO channel model in the presence of CCI. Space-time processing is used in order to deal with the frequency selectivity and to provide more degrees of freedom to deal with cochannel interference. We consider two non-linear space-time processing-based receivers. The first one is a MIMO space-time decision feedback equalizer (MIMO ST-DFE) and the second one is a space-time delayed decision feedback sequence estimator (MIMO ST-DDFSE) with actual adaptive algorithms for channel acquisition. Noise-limited as well as interference-limited situations are evaluated.

Parafac-based blind channel identification using 4th-order cumulants

Strong relationships between joint- diagonalization and tensor decompositions have been established recently. In this paper we propose a finite impulse response (FIR) channel identification method based on the Parafac decomposition of a 3rd-order tensor composed of the 4-th order output cumulants. By avoiding the prewhitening operation required by joint-diagonalization based methods, our method is shown to improve the estimation performance provided by the classic joint-diagonalization algorithm.

System-level evaluation of space-time processing for EDGE

We perform a system-level evaluation of conventional and decoupled space-time processing strategies for the enhanced data rates for global evolution (EDGE) system. The main idea of decoupled space-time (D-ST) processing is to enhance the performance of the temporal equalizer by separating cochannel interference (CCI) reduction from inter-symbol interference (ISI) suppression in two stages. Here, we employ a DDFSE as the temporal equalizer, which is a promising reduced-complexity equalization scheme for EDGE. In this contribution, the bit error rate (BER) performance of D-ST-DDFSE is accessed at both link- and system-levels. Concerning the system-level evaluation, an estimation of the output signal-to-interference-ratio (SIR) is performed for different load conditions and channel profiles such as the COST259 typical urban (TU) and bad urban (BU). Our objective is to show the potential gains of space-time processing in EDGE for extremely tight reuse patterns.

On supervised channel estimation techniques for very large MIMO communication systems

This paper compares and assesses the performance of different channel estimation techniques for multicell multiuser multiple-input multiple-output (MIMO) systems in a very large (VL) MIMO scenario. Some of these techniques exploit properties of large random matrices and are less affected by pilot contamination, which is the case of the technique based on the eingenvalue decomposition (EVD) of the output covariance matrix. Other techniques do not exploit such properties and are more sensitive to errors as the number of antenna sensors grow large, which is the case of the classical least squares (LS) method. Our main goal is to investigate the use of those techniques in a VL-MIMO system under different scenarios. This paper also proposes a new method to solve the multiplicative matrix ambiguity of the EVD-based method by using a simple Khatri-Rao product. Numerical results are shown to atest the increased effectiveness of the EVD-based channel estimation in a VL-MIMO environment.