A general detection framework for weak electrophysiological signals based on multi-channel flexible information fusion

Abstract

Existing decision-level information fusion methods for weak electrophysiological signals detection cannot effectively integrate conflicting information. Meanwhile, data-level information fusion methods lose physiological significance of channels. In this study, the evidence on each channel with a soft-decision method is first proposed. Then, a general multi-channel flexible information fusion detection framework for weak electrophysiological signals is established based on DSmT uncertain information fusion theory. Microvolt T wave alternans in ECG signal as an example, the proposed algorithm is verified on the simulation data and the measured data under various kinds of strong noise environments. These results demonstrate that the proposed method enables robust detection of weak electrocardiogram signals under strong noise background with significantly higher detection probability in the case of low false alarm probability by comparison with the existing decision-level fusion method, applying to the weak electrophysiological signals detection for mobile monitoring environments.