Low Order Approximation Distributed Sources Model and Direction-of-arrival Estimation

LI Qiang,LI Zhi-shun

DOI: https://doi.org/10.3969/j.issn.1002-8684.2007.12.015

2007-01-01

Abstract:The first order Taylor series expansion is used in the distributed sources model with generalized array manifold. Thus the model has no relationship with the form of the spatial distribution of the distributed sources and has more practicability. Here this model is called one order approximation distributed sources model. Research shows that the model error generated by neglecting the high order items of Taylor series is distinct. As a result, the performance of corresponding direction-of-arrival estimation algorithms will degrade. The two orders approximation distributed sources model with two orders Taylor series expansion is proposed. The new model can reduce the model error greatly and enhance the performance of direction-of-arrival estimation. Here the one order approximation model and two orders approximation model are called low order approximation model. Then the direction-of-arrival estimation algorithms with low order approximation model are analyzed and the generalized SMVDR algorithm is proposed. Computer simulations validate the performance of these algorithms. And the model error of low order approximation distributed sources model, spatial frequency distributed sources model and Jocabi-Anger series expansion distributed sources model are discussed.