S. Nakamori (Japan), A. Hermoso, J. Jiménez, and J. Linares (Spain)
Estimation. Statistical modelling. Chandrasekhar-type
filter. Uncertain observations. Covariance information.
In this paper, the least mean-squared error linear filtering
problem of a wide-sense stationary scalar signal from
uncertain observations is analyzed, assuming that the state
space model is not completely known. We suppose
that the variables modelling the uncertainty are not
independent; moreover, the observations are perturbed by
white plus coloured additive noises. We propose two
algorithms: one of them is based on Chandrasekhar-type
difference equations and, the other, on Riccati-type ones.
By comparing both algorithms it is observed that the
Chandrasekhar-type algorithm is computationally better
than the Riccati-type one since the number of difference
equations contained in it is less than that required in the
Riccati-type algorithm. Finally, we illustrate the results
obtained by means of a numerical example on estimation
of signals transmitted in multichannel.