At the beginning of my research studies, my research was focused on developing an efficient and robust algorithm to separate the cardiac and respiratory signals from an electrical bio-impedance signal. The separated signals could then be analysed by cardiologists to understand the conditions of the heart and lungs.

With respect to this problem, different approaches and methods, namely conventional filtering, independent component analysis, principle component analysis, wavelet, etc. were analysed, tested, and tried in order to solve the problem.

During the first part of my research, it has been understood what the mechanisms for the separation algorithms are and how to evaluate them. In particular, first part of research illustrated that it is not possible to evaluate the performance of the algorithms directly from measured data, because the parameters and waveform of the measured signals are uncertain, i.e. the cardiac and respiration signal parameters and waveform vary depending on the configuration of the electrodes, for example, and are subject to measurement errors.

The understanding and results of the first part of the research summarized above led to development of a signal model that can imitate the real phenomena of the cardiac and respiratory signals. The modelled signals could then be used for evaluating the performance of various signal processing algorithms, including separation algorithms.

Based on measured and cleaned extracted signals, the impedance cardiography (ICG) and impedance respirogram (IRG) signals have been modelled and a corresponding bio-impedance signal simulator (BISS) has been developed to simulate electrical bio-impedance (EBI) signals for evaluating the performance of various signal-processing algorithms on such signals.

In order to guide the development of the above signal models and simulator, a significant part of this research work focused on developing a physiological parametric framework for modelling measurable bioelectrical information and implement this parametric framework with a pragmatic approach on the bio-impedance example. Thus, in this study, a novel generic framework has been proposed for modelling the bioelectrical information and was then implemented for the case of EBI as an example.

From the author’s point of view, the following tasks and results have been conducted and achieved, respectively: