A cross-correlation technique as a system evaluation tool; application to blood flow measurement in extra-corporeal circuits

Auto-correlation and cross-correlation techniques have been favored by engineers and scientists as powerful analysis tools, especially when noise is present in the system. Since cross-correlation and power spectral density functions are tightly coupled, cross-correlation is a well suited instrument for the analysis of time invariant systems. This paper discusses the conditions under which cross-correlation techniques can be used to identify a system model. Random signals, either arbitrarily generated or inherent in the systems, can be used as test signals for the purpose of identification. A real time cross-correlator is designed and used to evaluate blood carrying vessels for the purpose of flow measurement in extra-corporeal circuits. Naturally occurring random signals due to 'impurities' in the blood such as red and white blood cells are correlated by means of two pairs of ultrasonic sensors. The correlation functions obtained under different flow rates manifest themselves as unit impulse responses of the fluid system between upstream and downstream sensors. The results obtained were evaluated for accurate modeling of the fluid path and transit time flow measurement.