The limitations of developing a technique to measure cardiac output continuously are given. Logical explanations are provided for the economic, technical, and physiologic benefits of a stochastic system identification technique for measuring cardiac output. Heat is supplied by a catheter-mounted filament driven according to a pseudorandom binary sequence. Volumetric fluid flow is derived by a cross-correlation algorithm written in the C language. In vitro validation is performed with water in a flow bench. The computed flow (y) compared with the in-line-measured flow (x) yields the linear regression y = 1.024x - 0.157 (r = 0.99). The average coefficient of variation is less than 2% over a volumetric fluid flow range of 2 to 10 L/min.