Experimental evaluation of DOA estimation of wideband LFM signals
Date
2019
Authors
Mulinde, R.
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 fulfillment of the requirements (Deliverable D1) of phase 3 of the collaborative research project conducted by the University of South Australia (UniSA) and the RF Systems Group of the Cyber and Electronic Warfare Division of the Defence Science and Technology (DST) group. It builds on and extends the results and analysis presented in previous reports (Deliverables D4 of Phase 1, D5 and D6 of Phase 2). Previous reports have investigated the performance of various direction of arrival (DOA) estimation techniques for wide-band signals mainly using synthetic data. Only a few preliminary results based on experimentally acquired data on linear frequency modulated (LFM) signals were presented in D6 (Phase 2 report). This report presents further results and analysis based on experimental LFM data using the channelized MUltiple SIgnal Classification (MUSIC) algorithm.
The reason for using channelized MUSIC is that this algorithm has relatively lower computational overhead as was shown in previous reports. DOA estimation performance for time-frequency MUSIC (T-F MUSIC) is briefly discussed.DOA estimation results are presented over a range of signal-to-noise ratios (SNRs), chirp rates, number of time segments, number of bins and true DOAs.The results for channelized MUSIC demonstrate that the DOA estimation performance for LFM signals improves as the chirp rate increases. Increasing the duration of the time-segments or equivalently reducing the number of time segments improves the DOA estimation performance. The DOA estimation performance deteriorates as the true DOA deviates from bore-sight.For channelized MUSIC, the DOA estimation performance improves with increasing SNR. However, for T-F MUSIC, the DOA estimation performance initially improves with SNR for the majority of chirp rates and then deteriorates (from SNR of 10 dB onwards). Based on the data available, only one exception to the above has been observed, namely, for the highest chirp rate of 50 MHz/µs, the DOA estimation performance does not seem to be much affected by the variation in SNR. However, this would warrant further investigation, perhaps using more experimental data points on both chirp rates and SNR.
School/Discipline
Dissertation Note
Provenance
Description
Data source: No, file://rdfs.unisa.edu.au/group_WBBF_algorithms/Archive_DoNot_Edit/Datasets/Phase3_D1/
Access Status
Rights
Copyright 2019 The Authors