Nihat Kabaoglu

Deterministic maximum likelihood approach for 3-D near field source localization

Summary In this paper, maximum likelihood estimator is proposed for passive localization of narrowband sources in the spherical coordinates (azimuth, elevation, and range). We adapt Expectation/Maximization iterative method to solve the complicated multi-parameter optimization problem appearing on the 3-D localization problem. The proposed algorithm is based on maximum likelihood criterion which employs the source signals recorded by 2-D array under near-field assumption. Expectation/Maximization algorithm decomposes the observed data into its components and then estimates the parameters of each signal component separately providing computationally efficient solution to the resulting optimization problem. Performance analysis of the proposed algorithm is then carried out through the evaluation of Cramer-Rao bounds. Finally, the applicability and effectiveness of the proposed algorithm is illustrated by some numerical simulations.

Target Tracking Using Particle Filters With Support Vector Regression

This paper presents a numerical Bayesian approach for the direction-of-arrival (DOA) tracking of multiple targets using a linear and passive sensor array. In this paper, support vector regression (SVR) method is employed, together with particle filters (PFs), to obtain an effective proposed distribution utilizing observed phenomena to propose a new sample. Two PF algorithms are presented: One is based on SVR for a large sample set, and the other is based on sequential SVR for a small sample set. The simulation results present the superiority of the proposed method while considering a small sample set and show that it is also competitive when a large sample set is considered.

Unconditional maximum likelihood approach for localization of near-field sources in 3D space

Since maximum likelihood (ML) approaches have better resolution performance than the conventional localization methods in the presence of less number and highly correlated source signal samples and low signal to noise ratios, we propose unconditional ML (UML) method for estimating azimuth, elevation and range parameters of near-field sources in 3D space in this paper. Besides these superiorities, stability, asymptotic unbiasedness, asymptotic minimum variance properties are motivated the application of ML approach. Despite these advantages, ML estimator has computational complexity. Fortunately, this problem can be tackled by the application of expectation/maximization (EM) iterative algorithm which converts the multidimensional search problem to one dimensional parallel search problems in order to prevent computational complexity.

Maximum likelihood 3-D near-field source localization using the EM algorithm

In this paper, maximum likelihood estimator is proposed for passive localization of narrowband sources in the spherical coordinates (azimuth, elevation, range). We adapt expectation/maximization iterative method to solve the complicated multi-parameter optimization problem appearing on the 3-D localization problem. The proposed algorithm is based on maximum likelihood criterion, which employs the source signals recorded by 2-D array under near-field assumption. Expectation/maximization algorithm decomposes the observed data into its components and then estimates the parameters of each signal component separately providing computationally efficient solution to the resulting optimization problem. Finally, some numerical simulations illustrate the applicability and effectiveness of the proposed algorithm.

SAGE Based Suboptimal Receiver for Downlink MC-CDMA Systems

An iterative joint data detection and channel estimation algorithm for downlink multi-carrier code division multiple access (MC-CDMA) systems is proposed. The resulting algorithm is a space alternating generalized expectation-maximization (SAGE) algorithm which updates the data sequences serially and the channel parameters in parallel, leading to a receiver structure that also incorparates interchannel interference cancellation. Its performance is compared with various Minimum Mean Square Error (MMSE) estimators for a multipath frequency selective fading channel using simulations. It is illustrated that the proposed SAGE algorithm performs better than the MMSE estimators considered in this paper.

Expectation–maximization-based joint data detection and channel estimation for a cellular multi-carrier code division multiple access network using single-hop relaying

SUMMARY This paper is concerned with channel estimation and data detection for a cellular multi-carrier code division multiple access network using single-hop relaying in the presence of frequency selective fading channels. The proposed expectation–maximization (EM) algorithm was used to jointly estimate both the coefficients of the channel between a relay and a base station and the data. EM algorithm is particularly suited to multi-carrier code division multiple access systems because they have multi-carrier signal format. The considered network uses single-hop relaying technique to provide a higher quality transmission to the users with low quality channels. The base station (managing mechanism) gives them an opportunity to send their messages via the users with high quality channels in a time sharing mode. The performance of the proposed EM algorithm, with and without hopping and with cooperative communication technique, was analyzed by a computer simulation, and the results are presented. Copyright

Wideband target tracking by using SVR-based sequential Monte Carlo method

In this work, a support vector regression (SVR) based sequential Monte Carlo method is presented to track wideband moving sources using a linear and passive sensor array for a signal model based on buffered data. The SVR method is employed together with a particle filter (PF) method to improve the PF tracker performance when a small sample set is available. SVR is used as a sample producing scheme for the current state vector. To provide a good approximation of the posterior density by means of improving the sample diversity, samples (particles) are drawn from an importance density function whose mean and covariance are calculated by using the pre-estimating state vector and the state vectors previous estimate. Thus, a better posterior density than the classical one can be obtained. Simulation results show that the method proposed in this work performs better than the classical one when a small sample set is available. Moreover, the results also show that a modified signal model that utilizes buffering data is superior to the signal model in Ng et al. [Application of particle filters for tracking moving receivers in wireless communication systems, in: IEEE Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Rome, Italy, June 2003, pp. 575-579].

EM-Based Recursive Tracking Algorithm for Near-Field Moving Sources

In this paper, we address the problem of joint tracking of the direction of arrival (DOA) and range parameters of moving sources in the near-field of an antenna array with the expectation-maximization (EM) based recursive algorithm. The main characteristic of the proposed recursive EM approach is to include computation of the gradient of the log-likelihood and some form of the complete-data fisher information matrix. The proposed recursive algorithm in this work assumes that the parameters of interest are described by a linear polynomial model. Simulation results of the suggested algorithm are also presented in order to illustrate the performance of the algorithms.

EM Based Stochastic Maximum Likelihood Approach for Localization of Near-field Sources in 3-D

Abstract The goal of this paper is to estimate the locations of unknown sources in 3-D space from the data collected by a 2-D rectangular array. Various studies employing different estimation methods under near-field and far-field assumptions were presented in the past. In most of the previous studies, location estimations of sources at the same plane with the antenna array were carried out by using algorithms having constraints for various situations indeed. In this study, location estimations of sources that are placed at a different plane from the antenna array is given. In other words, locations of sources in 3-D space is estimated by using a 2-D rectangular array. Maximum likelihood (ML) method is chosen as the estimator since it has a better resolution performance than the conventional methods in the presence of less number and highly correlated source signal samples and low signal to noise ratio. Besides these superiorities, stability, asymptotic unbiasedness, asymptotic minimum variance properties as well as no restrictions on the antenna array are motivated the application of ML approach. Despite these advantages, ML estimator has computational complexity. However, this problem is tackled by the application of Expectation/Maximization (EM) iterative algorithm which converts the multidimensional search problem to one dimensional parallel search problems in order to prevent computational complexity. EM iterative algorithm is therefore adapted to the localization problem by the data (complete data) assumed to arrive to the sensors separately instead of observed data (incomplete data). Furthermore, performance of the proposed algorithm is tested by deriving Cramer-Rao bounds based on the concentrated likelihood approach. Finally, the applicability and effectiveness of the proposed algorithm is illustrated by some numerical simulations

Support vector regression for surveillance purposes

This paper addresses the problem of applying powerful statistical pattern classification algorithm based on kernel functions to target tracking on surveillance systems. Rather than directly adapting a recognizer, we develop a localizer directly using the regression form of the Support Vector Machines (SVM). The proposed approach considers to use dynamic model together as feature vectors and makes the hyperplane and the support vectors follow the changes in these features. The performance of the tracker is demonstrated in a sensor network scenario with a constant velocity moving target on a plane for surveillance purpose.