|
Stepped-frequency waveforms are widely used in ultra-wideband (UWB) radar to obtain high resolution range profiles (HRRPs). In this paper, we consider high range resolution (HRR) improvement, i.e. sidelobe reduction and resolution enhancement, in stepped-frequency radar via weighted SParse Iterative Covariance-based Estimation (weighted SPICE) approach. Weighted SPICE is a unifying approach for four user parameter-free algorithms, namely SPICE, LIKES, SLIM and IAA. The latter three algorithms can be interpreted as weighted versions of SPICE with different data-dependent weights. Weighted SPICE is originally proposed for spectral estimation and array processing. We show that it can be used to obtain HRRPs in stepped-frequency radar as well, and comparisons among these four methods for HRR processing are also discussed in this paper. Additionally, different from spectral estimation applications, the main goal of HRR processing is to estimate the reflection coefficients {xk } of the targets rather than the powers {pk }. Thus, estimators which are able to obtain estimates {bxk } from {bpk }, such as Linear minimum mean square error (LMMSE) estimator and Capon estimator, are needed after obtaining {bpk } via weighted SPICE. As a minimum variance unbiased estimator, Capon is shown to have a better performance than LMMSE from the perspective of HRR processing. Numerical examples are presented to evaluate the performance of using all weighted SPICE algorithms with the LMMSE and Capon estimators, and IAA with Capon estimator is shown to outperform other weighted SPICE methods.
|
