Tansu Filik

2-D DOA estimation in case of unknown mutual coupling for multipath signals

In this paper, two dimensional (2-D) direction-of-arrival (DOA) estimation problem in case of unknown mutual coupling and multipath signals is investigated for antenna arrays. A new technique is proposed which uses a special array structure consisting of parallel uniform linear array (PULA). PULA structure is complemented with auxiliary antennas in order to have a structured mutual coupling matrix (MCM). MCM has a symmetric banded Toeplitz structure which allows the application of the ESPRIT algorithm for 2-D paired DOA estimation. The advantage of the PULA structure is exploited by dividing it into overlapping linear sub-arrays (triplets) and spatial smoothing is employed to mitigate multipath signals. Closed form expressions are presented for search-free, paired and unambiguous 2-D DOA estimation. Two algorithms PULA-1 and PULA-2 are proposed to effectively solve the problem. Several simulations are done and the accuracy of the proposed solution is shown.

A survey on the fundamentals of RSS based localization

In this research, for RSS based localization, geometry based new heuristic solutions are proposed and compared to maximum likelihood estimation (MLE) in terms of performance. It is shown that new heuristic methods which are geometrically meaningful can have almost equal performance with MLE. Furthermore, it is shown that they do not experience severe convergence problem unlike MLE. Therefore, during determination of cost functions, in addition to theoretical background, importance of the issues which are related to the numerical solution such as existences of local minima, overall structure and smoothness is emphasized.

GPS-based antenna tracking and signal beamforming system for small UAV platform

In this study GPS-based antenna tracking and signal beamforming system is considered to improve the data transmission quality, capacity and distance (range) of a small drone with ground station. The design phases and the performances for various cases are presented.

A joint detection and localization method for non-cooperative DS-SS signals

In this paper, a fast and computationally efficient joint presence detection and localization method is proposed for the non-cooperative direct sequence spread spectrum (DS-SS) signals. The method uses spatially distributed wideband sensor array output in a sparse way. It is applicable to long and timely varying spreading coded DS-SS signals. The proposed method uses the linearity of phase response between the cross elements of the spatial covariance matrices (R) along the signal bandwidth. This linearity in frequency domain is also used to estimate the time difference of arrival (TDOA) between sensors for the localization of these signals. A new sparse spatial covariance matrix (R) estimation technique which is called as block of consecutive frequency bins (BCFB) is also proposed. Hence, the computational complexity and data sharing rate between sensors are quite reduced according to the conventional incoherent signal subspace processing (ISSP) method. It is shown in simulations that the proposed BCFB method improves the TDOA estimation performance in low signal to noise ratio (SNR) and it attains to the Cramer Rao Lower Bound (CRLB) with greatly reduced processing time.

Presence detection of long-and-short-code DS-SS signals using the phase linearity of multichannel sensors

In this study, the detection of direct sequence spread spectrum (DS-SS) signals in white Gaussian noise without prior knowledge is investigated for spatially distributed multichannel wideband sensor array. A novel technique for the presence detection of both the long-and-short-code DS-SS signals is proposed which uses the frequency domain spatial covariance matrix of the sensor array. The technique is based on the linearity of the phase response of the wideband spatial covariances of cross sensors along the bandwidth of the signal. It is shown in simulations that the technique can detect DS-SS signals in a stable way at low signal to noise ratio (SNR) without any prior information and assumption on target. In addition, the proposed method can detect DS-SS signals in the presence of narrow band interference and multipath reflections.

A fast and automatically paired 2-dimensional direction-of-arrival estimation using arbitrary array geometry

A new approach is proposed for two-dimensional (2-D) direction-of-arrival (DOA) estimation with arbitrary array geometries, which is based on array interpolation. The method provides automatically paired source azimuth and elevation angle estimates. Furthermore it is possible to estimate D sources with D+1 sensor. 2-D array interpolation errors are minimized by using Wiener formulation. Proposed method is applied to the two planar arrays; uniform circular array (UCA) and uniform isotropic (IU) V-shaped array. It seen that the method can find closed form, automatically paired azimuth and elevation angles. Simulation results are presented verifying the efficacy of the method.

Direction of Arrival Estimation in Non-Uniform Linear Arrays for Wide Band Signals

Direction of arrival estimation of wideband signals takes an important place in both literature and practical applications. Although there are several studies on this topic, behavior of non-uniform linear arrays (NLA) for wideband signals was not investigated comprehensively. In this paper performance of Root-MUSIC algorithm is studied for wideband signals using NLA and a new approach is proposed. This new approach decreases the mean square error of the arrival angle estimation using Wiener formulation and initial estimate of the arrival angles. Advantages of NLA use for wideband signals are presented via several simulations.

Wind speed forecasting with missing values

In this study, a new short term wind speed fore-casting approach, which uses long term past observations, is proposed. The performance assessment of the wind speed forecasting framework is carried out using real data from meteorological stations in Marmara region of Turkey. The data sets are not complete due to equipment failures. The proposed approach builds on data de-trending, covariance-factorization via a subspace method, and one-step-ahead and multi-step-ahead Kalman filter prediction ideas. It is shown that trimming of diurnal, weekly, monthly, and annual patterns in data signify-cantly enhances estimation accuracy. Experimental test results demonstrate that the proposed multi-step-ahead forecasting out-performs the benchmark values computed with the persistent forecasting models.

EFFICIENCY ANALYSIS OF THE SOLAR TRACKING PV SYSTEMS IN ESKISEHIR REGION

In this study, efficiency analysis of the tracking solar photovoltaic (PV) system in Eskisehir region is considered for onsite real solar power generation values. It is shown that the tracking system’s power generations are always higher than fixed one, but increased yield ratio is changing according to weather conditions (sunny or cloudy times). In order to make a fair analysis/comparison, all the instantaneous real solar power generations between 1 July and 30 October 2016 are used to figure out the ratio of sunny and cloudy days of whole year. The averaged total generation values are calculated and it is revealed that tracking system provides approximately 33 % higher yield from same capacity fixed PV panels for Eskisehir region.

Detection of low probability of intercept signals in low SNR using multichannel sensor arrays

In this paper, a method is proposed for the detection of low probability of intercept, LPI signals in high noise without prior knowledge and assumption using spatially distributed multichannel wideband sensor array. The proposed detection method uses eigenvalue ratios (EVR) of wideband spatial covariance matrix as test statistic. It is shown in simulations that the proposed method can detect long code spread spectrum signals in a stable way even in low SNR. In addition, it is also shown in simulations that the proposed method is inherently resistive to the multipath reflections.