GCRIS

Now showing 1 - 3 of 3

Medium-Aware Inference for Wireless Sensor Networks
(Izmir Institute of Technology, 2020) Wahdan, Muath Abed Alrauf; Altınkaya, Mustafa Aziz
In a wireless sensor network, multilevel quantization is necessary in order to find a compromise between the smallest possible power consumption of the sensors and the detection performance at the fusion center (FC). The general methodology is using distance measures such as J-divergence (JD) and Bhattacharyya distance in this quantization. This thesis proposes a different approach which is based on maximizing the average output entropy of the sensors under both hypotheses of a binary hypothesis test and utilizes it in a Neyman-Pearson (NP) criterion based distributed detection scheme in order to detect a point source. Firstly, a deterministic signal and isotropic propagation model is considered. The receiver operating characteristics of the proposed maximum average entropy (MAE) meth\-od in quantizing sensor outputs was obtained for multilevel quantization both when the sensor outputs are available error-free at the FC and when non-coherent $M$-ary frequency shift keying communication is used for transmitting MAE based multilevel quantized sensor outputs over a Rayleigh fading channel. Secondly, the sequential testing version of the first problem is considered for both unquantized and quantized data transmissions. MAE and maximum JD (MJD) quantization methods for $M$-levels were applied in the sequential probability ratio test of Wald. The average sample number (ASN) required for the target probabilities of detection and a false alarm was the performance criterion: the smaller, the better. The performance of this test improves monotonically with the number of local sensors. Lastly, spatial correlation of the sensors is taken into the account. For this case, a Gaussian isotropic event source was applied. The computational requirements in evaluating multidimensional cumulative densities necessitated proposing a rectangular grid model of sensor deployment and block-diagonal approximations of covariance matrix related to the event signal at the sensors without losing generality. The simulation studies show the success of the MAE both in the cases of fusing error-free sensor outputs and in the case where the effect of the wireless channel is incorporated. As expected the performance gets better as the level of quantization increases and with six-level quantization, it approaches the performance of non-quantized data transmission. In the sequential tests again MAE was more successful compared to MJD resulting in smaller ASNs. It was observed that spatial correlation degrades system performance.
Generalized Bayesian model selection using reversible jump Markov chain Monte Carlo
(Izmir Institute of Technology, 2017) Karakuş, Oktay; Altınkaya, Mustafa Aziz; Kuruoğlu, Ercan Engin
The main objective of this thesis is to suggest a general Bayesian framework for model selection based on reversible jump Markov chain Monte Carlo (RJMCMC) algorithm. In particular, we aim to reveal the undiscovered potentials of RJMCMC in model selection applications by exploiting the original formulation to explore spaces of di erent classes or structures and thus, to show that RJMCMC o ers a wider interpretation than just being a trans-dimensional model selection algorithm. The general practice is to use RJMCMC in a trans-dimensional framework e.g. in model estimation studies of linear time series, such as AR and ARMA and mixture processes, etc. In this thesis, we propose a new interpretation on RJMCMC which reveals the undiscovered potentials of the algorithm. This new interpretation, firstly, extends the classical trans-dimensional approach to a much wider meaning by exploring the spaces of linear and nonlinear models in terms of the nonlinear (polynomial) time series models. Polynomial process modelling is followed by the definition of a new type of RJMCMC move that performs transitions between various generic model spaces irrespective of model sizes. Then, we apply this new framework to the identification of Volterra systems with an application of nonlinear channel estimation of an OFDM communication system. The proposed RJMCMC move has been adjusted to explore the spaces of di erent distribution families by matching the common properties of the model spaces such as norm, and this leads us to perform a distribution estimation study of the observed real-life data sets including, impulsive noise in power-line communications, seismic acceleration time series, remote sensing images, etc. Simulation results demonstrate the remarkable performance of the proposed method in nonlinearity degree estimation and in transitions between di erent classes of models. The proposed method uses RJMCMC in an unorthodox way and reveals its potential to be a general estimation method by performing the reversible jump mechanism between spaces of di erent model classes.
Inverse Problems and Regularization in Signal Processing With Applications To Wireless Channel Estimation
(Izmir Institute of Technology, 2011) Şahin, Ahmet; Altınkaya, Mustafa Aziz
The research presented in this thesis is on inverse problems encountered in the field of signal processing. Theory, classification and solution techniques of linear discrete inverse problems (LDIP) are investigated. LDIP are classified as underdetermined LDIP (ULDIP) and overdetermined LDIP (OLDIP). The solution methods developed for LDIP are applied to the particular problems of signal processing mainly channel estimation, equalization and compressive sampling. A new solution technique named constraint removal (CR) is presented for ULDIP type problems with sparse inputs. CR is applied to terrestrial digital TV (DTV) channel estimation. CR is also compared with subspace pursuit (SP) and linear programming. Regularization and optimum regularization parameter selection for ill-posed OLDIP type problems are discussed. Sparse channel estimation for wireless digital communications is investigated. A new channel estimation method, permuted deconvolution (PDEC), for long delay spread channels with short training sequences is proposed and compared with other methods. A review on equalization is presented. Different equalization techniques are discussed and compared. DFE is explained from an inverse problem perspective. A new non-feedback equalization technique called frequency compensated linear equalization (FC-LE) for sparse channels is presented and compared with DFE.

Phd Degree / Doktora

Browse

Filters

Settings

Sort By

Results per page

Search Results