Wideband Array Processing Algorithms for Acoustic Tracking of Ground Vehicles

Focused wideband adaptive array processing algorithms are used for the high-resolution aeroacoustic directionfinding of ground vehicles. Methods based on coherent focusing such as steered covariance matrix (STCM), and spatial smoothing or interpolation using the array manifold interpolation (AMI) method, will be discussed [1, 2]. Experimental results for a circular array are presented comparing and contrasting the coherent focused wideband methods with the incoherent wideband methods. Detailed analysis of the incoherent wideband MUSIC (IWM) algorithm versus the focused or coherent wideband MUSIC (CWM) algorithms shows processing gain over narrowband MUSIC in terms of accuracy and stability of the direction of arrival (DOA) estimates [3]. Given adequate SNR, incoherent wideband methods yield more accurate DOA estimates than focused methods for highly peaked spectra. On the other hand, the coherent wideband methods yield better DOA estimates for sources with flat spectra [4]. In general, the focused wideband methods are much more statistically stable than the incoherent wideband methods. In this paper, experimental results are shown for IWM, ICM, and AMI uniform circular array MUSIC (AMI-UCA MUSIC) algorithms using acoustic data collected from a 12element UCA and computational complexity is discussed.