Array Shape Estimation Using Partitioned Eigenstructure Method with Sources in Unknown Localizations

Advanced array processing approaches require accurate knowledge of the location of individual element in a sensor array. Most array shape estimation methods require the directions of sources. In this paper, an array shape estimation method based on eigen-decomposition is presented. The directions of sources do not need to be considered in advance and optimal array shape is generated through a series of iterations. To further improve the accuracy of this algorithm, a partitioned eigenstructure method is introduced. Numerical simulations using non-partitioned and partitioned method are conducted to verify the performance of the proposed technique.