A novel adaptive direction finding using Kalman algorithm
Y.-H. Chen, Ching‐Tai Chiang · 2002
In an attempt to improve the convergence rate, the authors introduce a novel Kalman noise-subspace estimator for the estimation of DOAs (directions of arrival). If the initial conditions are properly chosen as an identity matrix form, the Kalman-based estimator can estimate the complete noise subspace without a priori knowledge of the number of sources and the inflation method. The estimated weight vectors of the proposed algorithm are proved to approximately converge to the noise subspace for the high-SNR scenario. Simulation results show that the proposed algorithm has a much faster convergence rate and better mean square error performance than J. F. Yang and M. Kaveh's (1988) method.>