Detection by the stochastic classification matched filter
Nicolas Juennard, Bruno Borloz, Bernard Xerri · International Conference on Signal Processing · 2006
This paper deals with stochastic zero-mean signals detection. Second order statistical moments of signals is assumed to be known. We propose an adaptated method based on a linear filtering under constraint optimizing a criterion based on different signal to noise ratios. This optimisation leads to the choice of a 1D subspace on which the different signals are projected. We also present simulations and results obtained with different signals.