Biosignal Analysis and Interpretation: A toolbox for the biosignal processing

Abstract

This talk overviews methods of biosignal processing from the engineering point of view. Biosignals are categorized as either continuous or discrete processes. There are some cases where the same phenomenon could be processed as either a continuous or discrete process. If a signal is stationary, the family of joint probability distribution functions completely characterizes the signal. In practice, the joint moment functions of order up to four and their variants are commonly used for signal characterization. For nonstationary biosignals, adaptive signal processing techniques such as RLS or LMS algorithms are applied for the analysis. Signal decomposition by Kalman filter is also useful for the nonstationary analysis. For system identification based on observed signals, either linear or non-linear methods are utilized. This talk introduces useful signal analysis methods applicable to various kinds of biosignals, aiming to envision the rich field of biosignal interpretation and understanding.