Spatial analysis using new properties of the cross-spectral matrix
Jacques Munier, Gilles Y. Delisle · IEEE Transactions on Signal Processing · 1991
A novel concept, called the propagator, based upon a partition of the steering vectors matrix and, correlatively, of the received signals' cross-spectral matrix (CSM), is introduced for the characterization of source and noise subspaces without reference to the CSM eigenstructure. A localization technique can thereby be applied for uncorrelated or partially correlated sources with an arbitrary propagation model, which is directly comparable to the MUSIC algorithm. Given the signals produced by one particular source on some of the sensors, the propagator permits the extrapolation of the signals received by the remaining sensors. Various applications of this property can be found. An example is given concerning the estimation of the sensors' complex gains.>