Abstract: The problem of direction of arrival (DOA) estimation for multiple wideband sources, such as ground vehicles, using unattended passive acoustic sensors is considered in this thesis. Existing methods typically fail to detect and resolve DOAs of multiple closely spaced sources in tight formations, especially in the presence of interference and wind noise. This paper presents wideband extensions of several existing DOA estimation algorithms. The incoherently averaged MUSIC and WSF, which are signal subspace fitting algorithms, are applied and shown to have the best performance for the considered acoustic data sets. The Steered Covariance Matrix (STCM) algorithm is presented and applied to provide a benchmark of the capabilities of coherent wideband processing. Following frequency focusing, STCM uses Capon beamforming to provide high resolution DOA estimates. Different incoherent averages of the Capon spectrum across frequency bins are also explored and results are shown. Problems other than wind and interference such as sensor position error or wavefront perturbations caused by near-field effects or distributed sources occur in the data set used for testing these algorithms. An error specific to the randomly distributed wireless array is when data from an entire sensor channel is missed or received unusable, thus resulting in useless DOA estimates. The WSF and MUSIC algorithms show resilience to these errors. However, both of these algorithms are computationally expensive and modified versions of the wideband incoherently averaged Capon algorithm were sought to provide better performance at a smaller computational cost. A better understanding of the effects of these array uncertainties and source mismatches is sought so as to choose the correct algorithm for combating the encountered error. This part of the thesis includes a review of non-ideal signal models along with conclusions regarding how these models overlap. From an understanding of these models, appropriate robust algorithms are extended to wideband to combat error. A robust wideband Capon method is studied to account for some of these inherent problems in the array that can be caused by sensor position uncertainties and wavefront perturbations. Finally, to improve the resolution within a sector of interest and specifically to provide robustness to channel data loss, the beamspace method is extended and applied to the wideband problem. These methods are then tested and benchmarked on real acoustic signature data sets that contain multiple ground targets moving in various formations. Results show that better overall performance is obtained with the MUSIC and WSF algorithms, especially on the distributed array data set. However, particularly on the baseline uniform array data set, the wideband robust Capon and wideband beamspace Capon algorithms provide robust, accurate DOA estimation within the simple framework of the Capon algorithm.
Adviser: Dr. Mahmood R. Azimi-Sadjadi Co-Adviser: Non-ECE Member: Dr. F. Jay Breidt Member 3: Dr. Louis L. Scharf Addional Members:
