A Steering Vector Estimation Algorithm in Uncalibrated Arrays

The projection rotation scaling algorithm has perfect performance for estimating the steering vectors of uncalibrated isotropic sensors. But orders of the estimated steering vectors have randomicity in the algorithm. This paper presents a new steering vector initialization and signal subspace forming method. The method not only improves performance of the projection rotation scaling algorithm especially in low SNR conditions, but also can eliminate the randomicity of the orders. Numerical stimulations are provided to prove the effectiveness.