Analysis of DOA estimation performance for wideband LFM signals.
Date
2018
Authors
Mulinde, R.
Wenura Dissanayake, D.
Kaushik, M.
Ho, S.W.
Chan, T.
Attygalle, S.M.
Aziz, S.M.
Editors
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Report
Citation
Statement of Responsibility
Conference Name
Abstract
This report is in fulfilment of the requirement (Deliverable D1) of phase 2 of the collaborative research project conducted by University of South Australia(UniSA) and the RF Systems Group of the Cyber and Electronic Warfare Division of the Defence Science and Technology Group (DSTG). This report builds on and extends the results presented in the report submitted on the 29th of June 2018 (Deliverable D4 of phase 1). It is imperative that this report is read in conjunction with the previous report for a comprehensive understanding of the literature, the techniques and the results obtained.Several direction of arrival (DOA) estimation techniques for wideband nonstationary linear frequency modulated (LFM) signals that have been identified in phase 1 of this project are further analysed and compared in this phase. The three main categories of DOA estimation techniques investigated include time-frequency (T-F), fractional Fourier transform (FrFT) and adaptation of fast Fourier transform (FFT) techniques used in narrowband DOA estimation. In all these categories, DOA estimation is based on well-known narrowband MUltiple SIgnal Classification (MUSIC) and Estimation of Signal Parameters via Rotational Inverse Techniques (ESPRIT). The metrics used for performance evaluation include peak power, root mean-squared error (RMSE), half-power beamwidth (HPBW) and simulation time. Variation of DOA estimation performance with different parameters has been investigated. The parameters considered are received signal-to noise ratio (SNR), antenna array size, frequency resolution and sampling rate. The previous report investigated some of these variations. This report extends those investigations over a wider range of parameters. Furthermore, the variation of the RMSE with the listed parameters has been investigated.Additionally, DOA estimation performance of techniques based on ESPRIT has been added in this report. The RMSE for T-F is more susceptible to variation in sampling rate but less susceptible to variation in frequency resolution as compared to FrFT and FFT-based techniques. FFT-based narrowband approaches adapted for wideband DOA estimation exhibit smaller simulation time compared to both FrFT and T-F based techniques partly because they can be implemented using computationally efficient FFT routines.