Lower Bounds on Multiple-Source Direction Finding in the Presence of Direction-Dependent Antenna-Array-Calibration Errors.

摘要

We consider the problem of direction finding (DF) of multiple, cofrequency narrowband signals with a phased antenna array when direction-dependent array-calibration errors as well as thermal noise are present. Lower bounds on the variance of unbiased estimators for DF are derived under two different types of signal models: completely unknown (i.e., generic) signals and unknown constant-envelope signals. In both models, the complex amplitudes of the signals are modeled as unknown parameters. We derive and evaluate lower bounds on DF of these signals when complex Gaussian array errors and thermal noise are present. In our numerical examples, the bound for generic signals tended to decrease with increasing interference power and, for closely spaced signals, became more optimistic when small array errors were added. With a sufficiently large signal-of-interest (SOI) array signal-to-noise power ratio (ASNR) and number of looks, the bound numerically approached the bound on DF of multiple generic signals with array errors but no thermal noise (i.e., the multiple-generic-signal, array-errors-only bound). The results for constant-envelope signals showed that the bound with array errors and thermal noise was more optimistic than the analogous generic-signal bound and had little dependence on signal separation and interference power. (kt)