L<1) modeling), which is automaticallyP(z)H=h(0), where P(z)is the predictionHis the channel and its first coefficient. AlthoughP(z)of overestimated order, clever use ofHto channel length overestimation. This paper
identified by, e.g., a singular multichannel Levinson
algorithm, and can be shown to be robust to AR order
overestimation. On the other hand, K.A. Meraim and A.
Gorokhov derive other robustness properties based on the
equations
filter
using
the previous equations leads to robustness of the estimation
of
investigates these robustness issues, comparing both
methods (and derived methods) to order estimation algorithms
for, e.g., subspace-fitting methods. An important
point developed hereunder is the implicit order estimation
schemes present in linear prediction based methods
and their influence on identification performance. Furthermore,
we develop a new order estimation method, of
low computational cost and giving the channel estimate
as a by-product.
Linear prediction based algorithms have been applied to
the multi-channel FIR identification problem. In [11], it
was shown that oversampled and/or multiple antenna received
signals may be modeled as well as low rank MA
processes as low rank AR processes. Indeed, taking FIR
nature and the singularity of the MA process into account
(due to the fact that the number of channels is bigger than
the number of sources) leads to a finite order prediction
filter (i.e. AR(
