SUBSPACE SIGNAL-TO-NOISE RATIO MAXIMIZATION: THE CONSTRAINED STOCHASTIC MATCHED FILTER
Abstract
This paper deals with signal-to-noise ratio (SNR) max-imization in a subspace. This criterion can be applied to estimation or detection of a stochastic signals embedded in additive noise. Both signal and noise are considered to be realizations of random functions whose statistics are supposed known up to second-order. The method proposed leads to a filter we call " constrained stochastic matched filter " (CSMF), optimal in the sense that it maximizes the SNR in a subspace whose dimension is chosen a priori. It is an extension of the stochastic matched filter (SMF), itself an extension of the matched filter.
Domains
Signal and Image processing
Origin : Files produced by the author(s)
Loading...