Next Article in Journal
TSM-CV: Twitter Sentiment Analysis for COVID-19 Vaccines Using Deep Learning
Next Article in Special Issue
Moving Target Detection Algorithm for Millimeter Wave Radar Based on Keystone-2DFFT
Previous Article in Journal
A Study on Pavement Classification and Recognition Based on VGGNet-16 Transfer Learning
Previous Article in Special Issue
Water-Body Detection in Sentinel-1 SAR Images with DK-CO Network
 
 
Article
Peer-Review Record

Near-Field-to-Far-Field RCS Prediction Using Only Amplitude Estimation Technique Based on State Space Method

Electronics 2023, 12(15), 3371; https://doi.org/10.3390/electronics12153371
by Jinhai Huang 1,2, Jianjiang Zhou 1,* and Yao Deng 1
Reviewer 1: Anonymous
Reviewer 2:
Electronics 2023, 12(15), 3371; https://doi.org/10.3390/electronics12153371
Submission received: 10 July 2023 / Revised: 31 July 2023 / Accepted: 2 August 2023 / Published: 7 August 2023
(This article belongs to the Special Issue Advancements in Radar Signal Processing)

Round 1

Reviewer 1 Report

A brief summary:

Valuable approach for the transformation of the near-field (NF) to far-field (FF) RCS with adjusted amplitudes estimated using a state space approach for specific azimuth ranges.

General concept comments:

Do to dimensional constraints of e.g. anechoic chambers, it is only possible to measure the near-field the radar-cross-section (RCS) of a target under test, depending on the used frequency and size of the target itself. Therefore an approximation of the far-field RCS derived from the near-field is desired. The authors present an approach we firstly the measure near-field is corrected in amplitude and then transformed to the far-field. The authors also describe how to use the circular near-field-to-far-field transform (CNFFFT) to calculate the far-field on for a azimuth range instead of the full sphere.

The amplitude correction based on singular-value decomposition is a reasonable approach. Other possible approaches should at least be mentioned e.g. other eigen-based estimators or AR/ARMA models.

Further it would be nice to see a discussion about the problems resulting from the usage of the discrete Fourier transform for the (inverse)-Fourier-transforms needed for the CNFFFT.

Also a discussion about practical usage of the proposed approach should be including especially regarding problems that will arise e.g. calibration.

The result presentation is sufficient for the different distances and separate polarisations VV and HH.

Specific comments:

-There is a missing equation reference in line 255

-the legends in the result plots in Figure 10/11 could be right-oriented

The quality of the English language is sufficient.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report


Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Back to TopTop