Mustafa A. Altinkaya

A novel acoustic indoor localization system employing CDMA

Nowadays outdoor location systems have been used extensively in all fields of human life from military applications to daily life. However, these systems cannot operate in indoor applications. Hence, this paper considers a novel indoor location system that aims to locate an object within an accuracy of about 2 cm using ordinary and inexpensive off-the-shelf devices and that was designed and tested in an office room to evaluate its performance. In order to compute the distance between the transducers (speakers) and object to be localized (microphone), time-of-arrival measurements of acoustic signals consisting of Binary Phase Shift Keying modulated Gold sequences are performed. This DS-CDMA scheme assures accurate distance measurements and provides immunity to noise and interference. Two methods have been proposed for location estimation. The first method takes the average of four location estimates obtained by trilateration technique. In the second method, only a single robust position estimate is obtained using three distances while the least reliable fourth distance measurement is not taken into account. The system@?s performance is evaluated at positions from two height levels using system parameters determined by preliminary experiments. The precision distributions in the work area and the precision versus accuracy plots depict the system performance. The proposed system provides location estimates of better than 2 cm accuracy with 99% precision.

Subspace-based frequency estimation of sinusoidal signals in alpha-stable noise

In the frequency estimation of sinusoidal signals observed in impulsive noise environments, techniques based on Gaussian noise assumption are unsuccessful. One possible way to find better estimates is to model the noise as an alpha-stable process and to use the fractional lower order statistics (FLOS) of the data to estimate the signal parameters. In this work, we propose a FLOS-based statistical average, the generalized covariation coefficient (GCC). The GCCs of multiple sinusoids for unity moment order in SαS noise attain the same form as the covariance expressions of multiple sinusoids in white Gaussian noise. The subspace-based frequency estimators FLOS-multiple signal classification (MUSIC) and FLOS-Bartlett are applied to the GCC matrix of the data. On the other hand, we show that the multiple sinusoids in SαS noise can also be modeled as a stable autoregressive moving average process approximated by a higher order stable autoregressive (AR) process. Using the GCCs of the data, we obtain FLOS versions of Tufts-Kumaresan (TK) and minimum norm (MN) estimators, which are based on the AR model. The simulation results show that techniques employing lower order statistics are superior to their second-order statistics (SOS)-based counterparts, especially when the noise exhibits a strong impulsive attitude. Among the estimators, FLOS-MUSIC shows a robust performance. It behaves comparably to MUSIC in non-impulsive noise environments, and both in impulsive and non-impulsive high-resolution scenarios. Furthermore, it offers a significant advantage at relatively high levels of impulsive noise contamination for distantly located sinusoidal frequencies.

Estimation of the nonlinearity degree for polynomial autoregressiv processes with RJMCMC

Despite the popularity of linear process models in signal and image processing, various real life phenomena exhibit nonlinear characteristics. Compromising between the realistic and computationally heavy nonlinear models and the simplicity of linear estimation methods, linear in the parameters nonlinear models such as polynomial autoregressive (PAR) models have been accessible analytical tools for modelling such phenomena. In this work, we aim to demonstrate the potentials of Reversible Jump Markov Chain Monte Carlo (RJMCMC) which is a successful statistical tool in model dimension estimation in nonlinear process identification. We explore the capability of RJMCMC in jumping not only between spaces with different dimensions, but also between different classes of models. In particular, we demonstrate the success of RJMCMC in sampling in linear and nonlinear spaces of varying dimensions for the estimation of PAR processes.

Bayesian estimation of polynomial moving average models with unknown degree of nonlinearity

Various real world phenomena such as optical communication channels, power amplifiers and movement of sea vessels exhibit nonlinear characteristics. The nonlinearity degree of such systems is assumed to be known as a general intention. In this paper, we contribute to the literature with a Bayesian estimation method based on reversible jump Markov chain Monte Carlo (RJMCMC) for polynomial moving average (PMA) models. Our use of RJMCMC is novel and unique in the way of estimating both model memory and the nonlinearity degree. This offers greater flexibility to characterize the models which reflect different nonlinear characters of the measured data. In this study, we aim to demonstrate the potentials of RJMCMC in the identification for PMA models due to its potential of exploring nonlinear spaces of different degrees by sampling.

Long term wind speed prediction with polynomial autoregressive model

Wind energy is one of the preferred energy generation methods because wind is an important renewable energy source. Prediction of wind speed in a time period, is important due to the one-to-one relationship between wind speed and wind power. Due to the nonlinear character of the wind speed data, nonlinear methods are known to produce better results compared to linear time series methods like Autoregressive (AR), Autoregressive Moving Average (ARMA) in predicting in a period longer than 12 hours. A method is proposed to apply a 48-hour ahead wind speed prediction by using the past wind speed measurements of the Çeşme Peninsula. We proposed to model wind speed data with a Polynomial AR (PAR) model. Coefficients of the models are estimated via linear Least Squares (LS) method and up to 48 hours ahead wind speed prediction is calculated for different models. In conclusion, a better performance is observed for higher than 12-hour ahead wind speed predictions of wind speed data which is modelled with PAR model, than AR and ARMA models.

Frequency estimation of sinusoidal signals in alpha-stable noise using subspace techniques

In the frequency estimation of sinusoidal signals observed in impulsive noise environments, techniques based on Gaussian noise assumption are unsuccessful. One possible way to find better estimates is to model the noise as an alpha-stable process and to use the fractional lower order statistics of the data to estimate the signal parameters. Noise and signal subspace methods, namely the MUSIC and principal component-Bartlett methods, are applied to fractional lower order statistics of sinusoids embedded in alpha-stable noise. The simulation results show that techniques based on lower order statistics are superior to their second order statistics-based counterparts, especially when the noise exhibits a strong impulsive attitude.

Benefits of averaging lateration estimates obtained using overlapped subgroups of sensor data

In this paper, we suggest averaging lateration estimates obtained using overlapped subgroups of distance measurements as opposed to obtaining a single lateration estimate from all of the measurements directly if a redundant number of measurements are available. Least squares based closed form equations are used in the lateration. In the case of Gaussian measurement noise the performances are similar in general and for some subgroup sizes marginal gains are attained. Averaging laterations method becomes especially beneficial if the lateration estimates are classified as useful or not in the presence of outlier measurements whose distributions are modeled by a mixture of Gaussians (MOG) pdf. A new modified trimmed mean robust averager helps to regain the performance loss caused by the outliers. If the measurement noise is Gaussian, large subgroup sizes are preferable. On the contrary, in robust averaging small subgroup sizes are more effective for eliminating measurements highly contaminated with MOG noise. The effect of high-variance noise was almost totally eliminated when robust averaging of estimates is applied to QR decomposition based location estimator. The performance of this estimator is just 1 cm worse in root mean square error compared to the Cramer-Rao lower bound (CRLB) on the variance both for Gaussian and MOG noise cases. Theoretical CRLBs in the case of MOG noise are derived both for time of arrival and time difference of arrival measurement data.

Phase dependence mitigation for autocorrelation-based frequency estimation

The sinusoidal frequency estimation from short data records based on Toeplitz autocorrelation (AC) matrix estimates suffer from the dependence on the initial phases of the sinusoid(s). This effect becomes prominent when the impact of additive noise vanishes, that is at high signal-to-noise ratios (SNR). Based on both analytic derivation of the AC lag terms and simulation experiments we show that data windowing can mitigate the limitations caused by the phase dependence. Thus with proper windowing, the variance of the frequency estimate is no more eclipsed by phase dependence, but it continues to decrease linearly with increasing SNR. The study covers both the cases of a single sinusoid and two sinusoids closely spaced in the frequency with the Pisarenko frequency estimator, MUSIC and principal component autoregressive frequency estimators. The trade-offs between the spectral broadening and the achieved minimum variance level due to the data window are analyzed in detail.

Alpha-trimmed means of multiple location estimates

Localization by distance measurements is a common technique for solving this contemporary problem. The methods which achieve the theoretically optimum solutions have generally iterative structures. That is why when limited computational load is required, suboptimum methods described by closed form formulas like the one of Coope which depends on orthogonal decomposition of sensor coordinates, are preferred. In this method, when there are more than necessary distance measurements required for localization, the location will be found as the arithmetic average of the estimates obtained using the all three-combinations of distance measurements. In the averaging, eliminating the outlier estimates will increase the performance. In this case discarding the estimates making the ratio of alpha which are farthest away from the arithmetic average, one attains the so called alpha-trimmed mean of the estimates. Applying this technique, the disturbing effects of impulsive mixture of Gaussian contamination are eliminated and similar performances as in the case of Gaussian distance measurements are attained in localization.

Performance analysis of lattice reduction aided MIMO detectors

Lattice reduction is a powerful method used in detection and precoding of wireless multiple input-multiple output (MIMO) systems. The basic idea is to consider the channel transfer matrix as a basis for the transmitted symbols. The channel transfer matrix is reduced to a more orthogonal matrix using lattice reduction algorithms. This in turn, improves the performance of conventional MIMO receivers. In this study, it is shown that this performance improvement depends on the modulation order.