Graduate Exam Abstract

Amit Dutta

M.S. Final
June 17, 2020, 1:00 pm - 3:00 pm
Advanced spectral processing and analysis techniques for dual polarized weather radar

Abstract: This thesis focuses on the importance of spectral domain processing and analysis in weather radar applications such as sea-clutter mitigation and study of rain-hail mixtures in severe storms. An advanced spectral filtering technique has been proposed that helps in obtaining precipitation spectrum thus helping us to filter sea clutter and also closely study the spectrum of different rain and hail cases in severe storms.
Traditionally, time-domain auto-correlation techniques are used for estimation of dual-polarization radar moments from the raw IQ data. With the advent of low cost high-speed modern signal processors, frequency domain processing techniques are now feasible to be implemented in real-time. Hence spectral processing can be used for radar moments estimation. Previously, researchers have concluded that spectral filtering has improved the calculation of dual-pol radar moments. Many algorithms have been implemented in real-time for clutter mitigation and data quality control. In this thesis, various existing frequency and time domain algorithms such as standard notch filters, Gaussian Model Adaptive Processing (GMAP), and Parametric Time-Domain Method (PTDM) have been used for sea clutter mitigation and their performances are studied. Spectral Signal Quality Index (SSQI) which is dependent on the auto-correlation spectral density of the signals, has been used to threshold noisy spectrum to obtain a clean precipitation spectrum. Next using the results from PTDM along with the SSQI thresholding technique, the Adaptive Polarimetric Spectral Filter has been implemented for the clutter mitigation problem. The combination of these spectral filtering techniques is regarded as Advanced Adaptive Spectral Filter (AASF). The algorithms are applied to the CSU-SEAPOL radar data to identify and filter sea clutter. The AASF has been observed to perform better in terms of sea clutter suppression and identification.
In general, spectral analysis of radar time-series data reveals various characteristics of different hydrometeors. Incorporating Doppler information along with polarimetric measurements in dual-pol weather radar can unveil various microphysical properties in relation to the dynamics of storms in a radar resolution volume. This study is regarded as Spectral Polarimetry. Spectral analysis has been done on CSU-CHIVO radar data set which was collected during the Relampago campaign in Argentina. From the observations, spectral polarimetry revealed various features such as multi-modal spectrum, slopes in the spectral differential reflectivity, lowering of co-pol correlation spectrum, etc. These features essentially helped to characterize and determine the microphysical properties of different storms.
Thus the main goal of this thesis is to show the importance of spectral domain processing and analysis in relation to clutter mitigation and micro-physical study of storms.

Adviser: Dr. V Chandrasekar
Co-Adviser: N/A
Non-ECE Member: Dr. Thomas J Siller, Civil and Environmental Engineering
Member 3: Dr. Margaret Cheney, Electrical & Computer Engineering
Addional Members: N/A

A. Dutta and V. Chandrasekar, "Detection, Analysis and Mitigation of Sea Clutter in Polarimetric Weather Radar," 2019 United States National Committee of URSI National Radio Science Meeting (USNC-URSI NRSM), Boulder, CO, USA, 2019, pp. 1-2, doi: 10.23919/USNC-URSI-NRSM.2019.8712871.

Program of Study:
ECE 511
ECE 512
ECE 514
ECE 520
ECE 651
ECE 652
ECE 699