(kal’ man) A recursive filtering scheme applicable to linear systems, including time varying, nonstationary, and multichannel ones. A system is described by a model of first-order difference equations involving orthogonal state variables. The errors in measurements and the exciting disturbances are assumed to be Gaussian. In the non-Gaussian case, one can use an extended Kalman filter. The filter estimates the state variables based on prior measurements and the state-variable model, and incorporates the most recent measurements. The Kalman filter can be used as a recursive predictor. Kalman filtering is used in real-time reduction of integrated satellite-navigation data and in some seismic-filtering schemes, especially deconvolution. See Bayless and Brigham (1970) and Mendel and Kormylo (1978). Named for Rudolph Emil Kalman (1930-), Hungarian-American mathematician.