Lower bounds for wideband direction-finding with mutual coupling

This paper examines lower bounds on direction-of-arrival (DOA) estimation performance for radar systems with considerations made for wideband mutual coupling. For typical half-wavelength element spacing, array elements exhibit a degree of interaction with each other that may alter the currents on each element. We account for mutual coupling in the wideband case by implementing a least-squares inverse solution for mutual coupling interactions as a function of frequency. A coupling-impacted received data covariance is produced, which allows us to observe the effect of coupling on performance lower bounds, including the Cramér-Rao Lower Bound (CRLB) and the Ziv-Zakai Bound (ZZB). The ZZB was the focus as it holds at low SNR. Using the aforementioned bounds, we examined the impact of mutual coupling on different array sizes and compared dense to sparse array architectures. Additionally, we compared chirp waveforms to band-limited noise to compare bounds for different waveform types. We found that mutual coupling has a greater impact on the lower bound for smaller array sizes, and a sparse array configuration was found most favorable. The two waveform types demonstrate very similar behavior.