Sv. Schell et Wa. Gardner, PROGRAMMABLE CANONICAL CORRELATION-ANALYSIS - A FLEXIBLE FRAMEWORK FOR BLIND ADAPTIVE SPATIAL-FILTERING, IEEE transactions on signal processing, 43(12), 1995, pp. 2898-2908
We present a new framework known as the programmable canonical correla
tion analysis (PCCA) for the design of blind adaptive spatial filterin
g algorithms that attempt to separate one or more signals of interest
from unknown cochannel interference and noise. Unlike many alternative
s, PCCA does not require knowledge of the calibration data for the arr
ay, directions of arrival, training signals, or spatial autocorrelatio
n matrices of the the noise or interferers. A novel aspect of PCCA is
the ease with which new algorithms, targeted at capturing all signals
from particular classes of interest, can be developed within this fram
ework. In this paper, several existing algorithms are unified within t
he PCCA framework, and new algorithms are derived as examples. Analysi
s for the infinite-collect case and simulation for the finite-collect
case illustrate the operation of specific algorithms within the PCCA f
ramework.