Decomposition of the EBI Signal into Components using two Channel Cross-Compensating Singular Spectrum Analysis

Abstract

The paper describes a decomposition method of the Electrical Bioimpedance (EBI) signal into its components, in general, into cardiac and respiratory components respectively. Two-channel approach with cross-compensation is proposed in combination with the Singular Spectrum Analysis (SSA), which is applied for each channel individually. Moreover, a grouping algorithm of the SSA components is proposed, which is based on the cardiac signal periodicity and on the trajectory matrix features. The proposed algorithm is processing the EBI signal by frames of 9 seconds duration and can be extended for continuous EBI signal. The proposed algorithm makes possible to separate the cardiac and respiratory components of the total EBI signal in non-stationary conditions and with partially overlapping spectra of the cardiac and respiratory components.