Forum 04/07/2005

“Noise Artifact Minimization During Industrial Patient Monitoring Using System Theory Techniques”

Gail Baura, Vice President– Research, CardioDynamics International Corporation

The hospital environment is an infinite source of signal distortion. Electrosurgical interference, patient motion, respiration, blood pressure, and other noise sources make artifact minimization a necessary requirement before further digital signal processing and/or control can be applied to the signal of interest.  Under certain constraints of a periodic signal, a pseudorandom binary sequence filter may be used, which essentially functions as a bandpass filter that does not require frequency specifications. Alternatively, if the noise and signal frequency bands overlap, but a reference source for the noise is present, then an adaptive filter may be used to minimize the noise.  Finally, if the noise and signal frequency bands overlap, but the noise is minimal when transformed to another domain, then time-frequency or time-scale analysis may be used to minimize the noise.  These techniques, and their application to market-released patient monitors, will be discussed.