RESUMO

Adequate models of information signals and noises have been developed to solve urgent problems of synthesizing the hardware and software used for quick detection of survivors after man-made or natural disasters. It is shown that within the low-frequency range, comprising spectral components of the information signal from the human breathing and heartbeat, the adequate model of fluctuation noise is a flicker noise model built on the basis of recurrent equations of the Perron-Frobenius operator. The spectral characteristics of the information and noise signals are investigated. The information signal model built on the theory of periodically correlated random processes has been experimentally verified. An advanced signal processing algorithm using the process cross-sampling summing is proposed. A stochastic criterion for assessment of the signal duration period is the signal dispersion function against the lattice function period. The ergodic property of the process on the equidistant grid is theoretically proved and experimentally confirmed. By virtue of the criterion modality the optimum estimate of the period duration is the argument of maximal global extremum of this function. The simulation results are confirmed by the experimental data.