A modified TLS-Prony method using data decimation

Abstract

This paper introduces a modi ed TLS-Prony method which uses data decimation. The use of data decimation results in the reduction in the computational complexity in a variety of applications by allowing several low order estimations to be performed rather than one high order estimation. We also present an analysis of pole variance statistics for the modi ed TLS-Prony method which is used to explain and quantify the characteristics of decimation. We show that using decimation we can obtain comparable performance results at a fraction of the computational cost.

Attributed scattering centers for SAR ATR

“…Statistical analysis of both 1-D [47], [48], [39], [49] and 2-D [22], [43], [50] damped exponential estimators, and corresponding CR bounds [43], [51], are available. The "true 2-D" methods generally outperform the combined 1-D methods because they exploit the damped exponential structure in both dimensions simultaneously.…”

“…In particular, we consider both damped and undamped exponential signal models, which have been used extensively for radar signal analysis [11], [24], [2], [3], [37], [22], [17]. For each of these reduced complexity models, we present the parametric form, relate the parameters to the scattering center attributes parameterized by GTDbased model in (8), bound the approximation error, review algorithms for estimating the model parameters, and highlight parameter estimation issues arising from use of the surrogate models.…”

NMR Data Analysis: A Time-Domain Parametric Approach Using Adaptive Subband Decomposition

“…The subband approach may be seen as a pre-processing of the signal, and thus is independent from the method used for the estimation. In addition to their computational efficiency, these techniques present several other advantages comparatively to a fullband estimation [32,36,37]. In particular, it is known that data decimation (or downsampling) may increase the resolution capabilities of the estimator considered [38] and enables one to process signals at low signal-to-noise ratios [39].…”

“…The last two aforementioned problems have motivated, last years, a renewed interest for subband-based approaches [30][31][32][33][34] that have been known for a long time in the signal processing community. The idea is to process a signal by splitting it into several small spectral windows.…”

“…It can be shown that the resulting signals can still be modeled as a combination of (a reduced number of) modes [32,44]. For notation simplicity, let x (n) be a subband signal and v be the subband image of the fullband parameter v, then the model of the subband signal is:…”