IEEE signal processing letters

Vol. 26, num. 7, p. 1085-1089

DOI: 10.1109/LSP.2019.2918945

Date of publication: 2019-07-01

Abstract:

Decision-making procedures when a set of individual binary labels is processed to produce a unique joint decision can be approached modeling the individual labels as multivariate independent Bernoulli random variables. This probabilistic model allows an unsupervised solution using EM-based algorithms, which basically estimate the distribution model parameters and take a joint decision using a Maximum a Posteriori criterion. These methods usually assume that individual decision agents are conditionally independent, an assumption that might not hold in practical setups. Therefore, in this work we formulate and solve the decision-making problem using an EM-based approach but assuming correlated decision agents. Improved performance is obtained on synthetic and real datasets, compared to classical and state-of-the-art algorithms.

© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.]]>

IEEE signal processing letters

Vol. 20, num. 6, p. 595-598

DOI: 10.1109/LSP.2013.2260329

Date of publication: 2013-01

Abstract:

We address the problem of distributed estimation of a parameter from a set of noisy observations collected by a sensor network, assuming that some sensors may be subject to data failures and report only noise. In such scenario, simple schemes such as the Best Linear Unbiased Estimator result in an error floor in moderate and high signal-to-noise ratio (SNR), whereas previously proposed methods based on hard decisions on data failure events degrade as the SNR decreases. Aiming at optimal performance within the whole range of SNRs, we adopt a Maximum Likelihood framework based on the Expectation-Maximization (EM) algorithm. The statistical model and the iterative nature of the EM method allow for a diffusion-based distributed implementation, whereby the information propagation is embedded in the iterative update of the parameters. Numerical examples show that the proposed algorithm practically attains the Cramer–Rao Lower Bound at all SNR values and compares favorably with other approaches.]]>

IEEE signal processing letters

Vol. 14, num. 5, p. 359-362

Date of publication: 2007-05

IEEE signal processing letters

Vol. 81, num. 8, p. 1581-1602

Date of publication: 2001-08

IEEE signal processing letters

Vol. 81, num. 8, p. 668-672

DOI: 10.1109/SSAP.2000.870210

Date of publication: 2000-08

Abstract:

The estimation of the delay of a known training signal received by an antenna array in a multipath channel is addressed. The effect of the co-channel interference is taken into account by including a term with unknown spatial correlation. The channel is modeled as an unstructured FIR filter. The exact maximum likelihood (ML) solution for this problem is derived, but it does not have a simple dependence on the delay. An approximate estimator that is asymptotically equivalent to the exact one is presented. Using an appropriate reparameterization, it is shown that the delay estimate is obtained by rooting a low-order polynomial, which may be of interest in applications where fast feedforward synchronization is needed.]]>

IEEE signal processing letters

Vol. 4, num. 7, p. 207-209

DOI: 10.1109/97.596889

Date of publication: 1997-07

Abstract:

A discussion on the expression proposed by Weiss et al. (see J. Acoust. Soc. Amer., vol.96, p.850-6 and p.857-66, 1994 and IEEE Signal Processing Mag., vol.11, p.13-32, 1994) for deconvolving the wideband density function is presented. We prove here that such an expression reduces to be proportional to the wideband correlation receiver output, or continuous wavelet transform of the received signal with respect to the transmitted one. Moreover, we show that the same result has been implicitly assumed by Weiss et al., when the deconvolution equation is derived. We stress the fact that the analyzed approach is just the orthogonal projection of the density function onto the image of the wavelet transform with respect to the transmitted signal. Consequently, the approach can be considered a good representation of the density function only under the prior knowledge that the density function belongs to such a subspace. The choice of the transmitted signal is thus crucial to this approach.

A discussion on the expression proposed in [1]–[3] for deconvolving the wideband density function is presented. We prove here that such an expression reduces to be proportional to the wideband correlation receiver output, or continuous wavelet transform of the received signal with respect to the transmitted one. Moreover, we show that the same result has been implicitly assumed in [1], when the deconvolution equation is derived. We stress the fact that the analyzed approach is just the orthogonal projection of the density function onto the image of the wavelet transform with respect to the transmitted signal. Consequently, the approach can be considered a good representation of the density function only under the prior knowledge that the density function belongs to such a subspace. The choice of the transmitted signal is thus crucial to this approach.]]>