A new approach for determining the spectral data of multichannel harmonic signals in noise (Q909637)

This paper is concerned with detecting and estimating the frequencies and amplitude matrices associated with multichannel harmonic signals in the presence of additive noise. The Wold decomposition and the multichannel Levinson Algorithm are used to provide the basis for a recursive procedure identifying the signal frequencies and amplitude matrices. The results obtained are a generalization of previous work concerning the single channel setting. The paper provides a theoretical basis for the multichannel estimation problem. Computational/algorithmic issues require further attention.