João Cesar M. Mota

PARAFAC-based unified tensor modeling for wireless communication systems with application to blind multiuser equalization

In some antenna array-based wireless communication systems the received signal is multidimensional and can be treated as a tensor (3D array) instead of a matrix (2D array). In this paper, we make use of a generalized tensor decomposition known as constrained Block-PARAFAC and propose a tensor (3D) model for the signal received by three types of wireless communication systems. The considered wireless communication systems are multiuser systems subject to frequency-selective multipath and employing multiple receiver antennas together with (i) oversampling or (ii) direct-sequence spreading or (iii) multicarrier modulation. The proposed modeling approach aims at unifying the received signal model of these systems into a single PARAFAC model. We show that the proposed model has a constrained structure, where model constraints and associated dimensions depend on each particular system. The proposed constrained Block-PARAFAC model is demonstrated by expanding the tensor using Kronecker products of canonical vectors. As an application of this model to tensor signal processing, a new tensor-based receiver is proposed for blind multiuser equalization, which combines PARAFAC-based modeling with a subspace method. Simulation results are presented to illustrate the performance of the proposed blind receiver.

A Constrained Factor Decomposition With Application to MIMO Antenna Systems

In this paper, we formulate a new tensor decomposition herein called constrained factor (CONFAC) decomposition. It consists in decomposing a third-order tensor into a triple sum of rank-one tensor factors, where interactions involving the components of different tensor factors are allowed. The interaction pattern is controlled by three constraint matrices the columns of which are canonical vectors. Each constraint matrix is associated with a given dimension, or mode, of the tensor. The explicit use of these constraint matrices provides degrees of freedom to the CONFAC decomposition for modeling tensor signals with constrained structures which cannot be handled with the standard parallel factor (PARAFAC) decomposition. The uniqueness of this decomposition is discussed and an application to multiple-input multiple-output (MIMO) antenna systems is presented. A new transmission structure is proposed, the core of which consists of a precoder tensor decomposed as a function of the CONFAC constraint matrices. By adjusting the precoder constraint matrices, we can control the allocation of data streams and spreading codes to transmit antennas. Based on a CONFAC model of the received signal, blind symbol/code/channel recovery is possible using the alternating least squares algorithm. For illustrating this application, we evaluate the bit-error-rate (BER) performance for some configurations of the precoder constraint matrices.

Condition monitoring of 3G cellular networks through competitive neural models

We develop an unsupervised approach to condition monitoring of cellular networks using competitive neural algorithms. Training is carried out with state vectors representing the normal functioning of a simulated CDMA2000 network. Once training is completed, global and local normality profiles (NPs) are built from the distribution of quantization errors of the training state vectors and their components, respectively. The global NP is used to evaluate the overall condition of the cellular system. If abnormal behavior is detected, local NPs are used in a component-wise fashion to find abnormal state variables. Anomaly detection tests are performed via percentile-based confidence intervals computed over the global and local NPs. We compared the performance of four competitive algorithms [winner-take-all (WTA), frequency-sensitive competitive learning (FSCL), self-organizing map (SOM), and neural-gas algorithm (NGA)] and the results suggest that the joint use of global and local NPs is more efficient and more robust than current single-threshold methods.

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.

ST-segment analysis using hidden Markov Model beat segmentation: application to ischemia detection

In this work, we propose an ECG analysis system to ischemia detection. This system is based on an original markovian approach for online beat detection and segmentation, providing a precise localization of all beat waves and particularly of the PQ and ST segments. Our approach addresses a large panel of topics never studied before in others HMM related works: multichannel beat detection and segmentation, waveform models and unsupervised patient adaptation. Thanks to the use of some heuristic rules defined by cardiologists, our system performs a reliable ischemic episode detection, showing to be a helpful tool to ambulatory ECG analysis. The performance was evaluated on the two-channel European ST-T database, following its ST episode definitions. The experimentation was performed over 48 files extracted from 90. Our best average statistic results are 83% sensitivity and 85% positive predictivity. Performance compares favorably to others reported in the literature.

Space-Time Multiplexing Codes: A Tensor Modeling Approach

In this paper, we present new space-time multiplexing codes (STMC) for multiple-antenna transmissions, which rely on a three-dimensional tensor modeling of the transmitted/received signals. The proposed codes combine spatial multiplexing and space-time coding by spreading a linear combination of different sub-streams of data over the space and time dimensions. We show the STMC induces a tensor structure on the transmitted/received signal that can be modeled using a trilinear tensor decomposition. Tensor modeling is exploited at the receiver for a blind decoding of the transmitted sub-streams based on linear processing and without any ambiguity. The proposed approach also provides full diversity while benefiting from the maximum multiplexing gain offered by the multiple antennas. Simulation results show that the tensor-based STMC offer remarkable performance with good diversity-multiplexing trade-off

Space-time spreading-multiplexing for MIMO wireless communication systems using the PARATUCK-2 tensor model

In this paper, we present a new space-time spreading-multiplexing model for multiple-input multiple-output (MIMO) wireless communication systems relying on a tensor modeling of the transmitted and received signals. At the transmitter, we exploit the core of a PARATUCK-2 tensor model composed of a precoding matrix and two allocation matrices that allow to control the spreading and multiplexing of the data streams across the space dimension (transmit antennas) and time-dimension (time-slots). Different MIMO schemes combining space-time multiplexing and diversity can be derived from the proposed model. The identifiability and uniqueness of the PARATUCK-2 tensor model for the received signal are discussed and subsequently exploited for a joint blind channel estimation and symbol detection. The bit-error-rate performance of different transmit schemes derived from the proposed tensor model is evaluated by means of computer simulations.

Decision directed adaptive blind equalization based on the constant modulus algorithm

In this paper, new decision directed algorithms for blind equalization of communication channels are presented. These algorithms use informations about the last decided symbol to improve the performance of the constant modulus algorithm (CMA). The main proposed technique, the so called decision directed modulus algorithm (DDMA), extends the CMA to non-CM modulations. Assuming correct decisions, it is proved that the decision directed modulus (DDM) cost function has no local minima in the combined channel-equalizer system impulse response. Additionally, a relationship between the Wiener and DDM minima is established. The other proposed algorithms can be viewed as modifications of the DDMA. They are divided into two families: stochastic gradient algorithms and recursive least squares (RLS) algorithms. Simulation results allow to compare the performance of the proposed algorithms and to conclude that they outperform well-known methods.

Tensor-Based Space-Time Multiplexing Codes for MIMO-OFDM Systems with Blind Detection

A new approach to space-time-frequency coding for multiple-input multiple-output (MIMO) systems based on orthogonal frequency division multiplexing (OFDM) is presented. Tensor-based space-time-multiplexing (TSTM) codes combine multi-stream spatial multiplexing and transmit diversity, and are based on a tensor modeling of the transmitted/received signals. The proposed codes are designed to offer some transmission flexibility by allowing a simple multiplexing-diversity-rate control as well as to achieve full space and multipath diversities in a frequency-selective channel. We show that the received signal has a tensor structure and this tensor modeling is exploited for blind separation/decoding of the transmitted information. Simulation results illustrate the performance of some TSTM codes with blind detection

Multiuser MIMO system using block space-time spreading and tensor modeling

In this paper, we consider a point-to-multipoint downlink multiuser wireless communication system, where a multiple-antenna base station simultaneously transmits data to several users equipped with multiple receive antennas. The transmit antenna array is partitioned into transmission blocks, each one being associated with a given user. Space-time spreading is performed within each block using a transmit antenna subset. We formulate block space-time spreading using a tensor modeling. We show that the tensor-based block space-time spreading model has the distinguishing feature of modeling a multiuser space-time transmission with different spatial spreading factors (diversity gains) as well as different multiplexing factors (code rates) for the users. The space-time spreading structure is chosen to allow a deterministic multiuser interference (MUI) elimination by each user. A block-constrained tensor model is then presented for the received signal, which is characterized by fixed constraint matrices that reveal the overall space-time spreading pattern. At each receiver, blind joint channel and symbol recovery is performed using an alternating least squares algorithm. Simulation results illustrate the performance of the proposed transceiver model in terms of bit-error-rate, channel/symbol estimation accuracy and link-level throughput.