Optimum eigenfilters and matched filters

The authors propose three criteria for maximising the signal-to-noise ratio in detecting a signal in noise. In the stochastic case, the optimal filter maximising the average power of the signal to the average power of the noise is an eigenfilter, and is usually considered as the stochastic counterpart of the matched filter. This interpretation is inexact. However, the optimum eigenfilter can be viewed as a filter matched to the principal component of the signal´s power spectral density. A fast, efficient technique for the computation of the optimum eigenfilter is proposed. Simulation results show the efficiency of the proposed algorithm