INSTRUMENTAL VARIABLE APPROACH TO ARRAY-PROCESSING IN SPATIALLY CORRELATED NOISE FIELDS

High-performance signal parameter estimation from sensor array data is a problem which has received much attention. A number of so-called ei genvector (EV) techniques such as MUSIC, ESPRIT, WSF, and MODE have be en proposed in the literature. The EV techniques for array processing require knowledge of the spatial noise correlation matrix that constit utes a significant drawback. Herein, a novel instrumental variable (IV ) approach to the sensor array problem is proposed. The IV technique r elies on the same basic geometric properties as the EV methods to obta in parameter estimates. However, by exploiting the temporal con elated ness of the source signals, no knowledge of the spatial noise covarian ce is required. The asymptotic properties of the IV estimator are exam ined and an optimal IV method is derived. Computer simulations are pre sented to study the properties of the IV estimators in samples of prac tical length. The proposed algorithm is also shown to perform better t han MUSIC on a full-scale passive sonar experiment.