A Source Representation Space Approach To Eigenstructure Based Broadband Source Location Estimation

For broadband Source Location Estimation (SLE), two new broadband spatial spectral estimation algorithms are described which utilize the eigenstructure of a broadband spatial/temporal data covariance matrix. The algorithms are formulated using a broadband source model, termed the broadband source representation space. The model, a subspace of the source observation space, is developed and employed to establish the plausibility of broadband matrix eigenstructure methods. For one of the proposed algorithms, broadband source representation space basis vectors are used directly. Simulations are presented illustrating performance and extensions are noted.