Direction of Arrival (DOA)Estimation based on Sparsity-Aware Signal Processing

Please enable javascript in your browser.

Nikoomahasen Sarukolaee, Ahmad | 2020

328 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 52898 (05)
University: Sharif University of Technology
Department: Electrical Engineering
Advisor(s): Behnia, Fereidoon; Babaiezadeh, Massoud
Abstract:
Estimating direction of arrival (DOA) is one of the most important problems in array signal processing to solve which various methods have been proposed. The older methods for estimating signal DOA were divided into three main groups: beamforming, maximum likelihood-based and subspace-based methods. By applying sparse representation techniques to the DOA estimation problem, a new group of methods for solving this problem are introduced. In this thesis, two grid-based methods, which are tow sub groups of sparse methods for estimation of DOA, are proposed. Each of these methods uses singular value decomposition to reduce the power of noise. Also proposed methods are compared with the multiple signal classification method and the l_(2,1) norm-based methods in terms of resolution and accuracy. Finally, by converting the direction of arrival estimation problem into a classification problem and using a convolutional neural network, the DOA is estimated. The simulation results show high resolution and accuracy of this method in low noise to signal ratio scenarios
Keywords:
Direction of Arrival (DOA)Estimation ; Sparse Representation ; Deep Learning ; Convolutional Neural Network ; Sparse On-Grid Method ; Toeplitz Covariance Matrix

No TOC