Digital Techniques for the Approximate Identification and Simulation of the Respiration: Heartrate System

摘要

The problem of determining a digital computer model for relating heart rate variations to respiration and other nonrandom factors is considered. The respiratory signal is a measured signal while the other nonrandom factors are unmeasurable, the sum of which is called a nonrespiratory signal. The model is derived from heart beat times and sampled values of the respiratory signal. A process is developed whereby the heart beat times are converted to a uniformly sampled heart rate signal. This allows a sampled-data model to be used to approximate the respiration-heartrate system which is assumed to be continuous. The respiratory signal is the system input and the heart rate signal is considered to be the sum of the system output, the nonrespiratory signal and noise. Several identification methods are considered for determining the parameters of the respiration-heartrate system transfer function. An on-line type identification method using instrumental variables is developed so that indefinite lengths of data may be used. Experimental data from several subjects are processed to illustrate the usefulness of the methods developed. (Author)