High resolution adaptive bearing estimation using a complex-weighted neural network

• Published in: [Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing Vol. 2; pp. 317 - 320 vol.2
• Main Authors: Yupeng Chen, Chaohuan Hou
• Format: Conference Proceeding
• Language: English
• Published: IEEE 1992
• Subjects: Chaos; Data mining; Direction of arrival estimation; Narrowband; Neural networks; Neurons; Personal communication networks; Phased arrays; Principal component analysis; Signal resolution
• ISBN: 9780780305328; 0780305329
• ISSN: 1520-6149
• DOI: 10.1109/ICASSP.1992.226056

A neuron-based algorithm for solving the complex principal components analysis problem and its application to adaptive bearing estimation are presented. The authors specify the bearing estimation problem in a narrowband version and use the eigen-decomposition method to achieve high resolution. Both input data and eigenvectors that span the signal subspace are complex values. So it is important to extract the complex principal components from the complex input data sequence. Previous methods do not offer complex algorithms. To overcome this problem, the authors propose a linear neural network with complex weights which is a generalized and modified version of E. Oja's (1985) and S.Y. Kung and K.I. Diamantaras's (1990) methods, and they use their own method to estimate the direction of arrival (DOA).< >