Showing papers by "Mats Viberg published in 1997"

Maximum likelihood array processing in spatially correlated noise fields using parameterized signals

Abstract: This paper deals with the problem of estimating signal parameters using an array of sensors. This problem is of interest in a variety of applications, such as radar and sonar source localization. A vast number of estimation techniques have been proposed in the literature during the past two decades. Most of these can deliver consistent estimates only if the covariance matrix of the background noise is known. In many applications, the aforementioned assumption is unrealistic. Recently, a number of contributions have addressed the problem of signal parameter estimation in unknown noise environments based on various assumptions on the noise. Herein, a different approach is taken. We assume instead that the signals are partially known. The received signals are modeled as linear combinations of certain known basis functions. The exact maximum likelihood (ML) estimator for the problem at hand is derived, as well as computationally more attractive approximation. The Cramer-Rao lower bound (CRB) on the estimation error variance is also derived and found to coincide with the CRB, assuming an arbitrary deterministic model and known noise covariance.

Separable non-linear least-squares minimization-possible improvements for neural net fitting

Abstract: Neural network minimization problems are often ill-conditioned and in this contribution two ways to handle this will be discussed. It is shown that a better conditioned minimization problem can be obtained if the problem is separated with respect to the linear parameters. This will increase the convergence speed of the minimization. The Levenberg-Marquardt minimization method is often concluded to perform better than the Gauss-Newton and the steepest descent methods on neural network minimization problems. The reason for this is investigated and it is shown that the Levenberg-Marquardt method divides the parameters into two subsets. For one subset the convergence is almost quadratic like that of the Gauss-Newton method, and on the other subset the parameters do hardly converge at all. In this way a fast convergence among the important parameters is obtained.

Two decades of statistical array processing

Abstract: Array signal processing is the common name for a large number of signal processing techniques involving parameter estimation from multichannel data. The prototype problem is to find the directions of incoming wavefronts using an antenna array. The applications are numerous, including unexpected problems not involving spatially distributed sensors. A vast number of estimation methods have been proposed and extensively analyzed over the last 2-3 decades. The paper provides an introduction to the various algorithms.

Performance of decoupled direction finding based on blind signal separation

Abstract: A number of methods for blindly separating superimposed digitally modulated signals arriving at an antenna array have been proposed. These techniques are efficient at the up-link (mobile to base) in a mobile communication system. However, for solving the base-to-mobile beamforming problem it may be necessary to also estimate the directions-of-arrival (DOAs) of the various signal paths. We present an optimal decoupled DOA estimation procedure based on information from the blind separation algorithm. Its performance is evaluated in the presence of spatially correlated noise and array modeling errors. The proposed technique has computational advantages as compared to traditional DOA estimation, because the different signal waveforms are treated in a separated fashion. Yet, the decoupled approach is shown to be substantially less sensitive to modeling errors and interference.

Bayesian Approaches for Robust Array Signal Processing

Abstract: Modern techniques for sensor array signal processing are hampered by the often un realistic and overly simplistic assumptions used in their development Foremost among these assumptions are those relating to the array geometry or response including for example known sensor positions availability of complete gain phase mutual coupling calibration data uniform linear arrays etc Of course with real arrays none of these assumptions hold exactly Deviations from the nominal model occur due to environmen tal factors quantization e ects perturbations in the locations of the antenna elements etc If such model errors are ignored serious performance degradation can result While in any given application the exact value of the perturbation to the array is unknown the size or distribution of such perturbations may be well understood Consequently in this paper a Bayesian approach is adopted to show how information in the form of an a priori distribution on the array model errors can be used to improve both direction of arrival and beamforming performance The general maximum a posteriori estimator for the problem is formulated and a computationally attractive alternative based on the concept of subspace tting is proposed The algorithm s statistical performance for both DOA estimation and beamforming is evaluated by means of some simulation examples