Impulse control of observations in infinite-dimensional Kalman filtering
Michael Basin · 1999
Develops the impulse control approach to the observation process in infinite-dimensional Kalman filtering problems, that results in introducing discrete observations into a continuous observation process. Impulsive modeling of the transition operator in an observation equation generates online computable jumps of the estimate correlation function from its current position towards zero and, as a result, leads to instantaneous reduction of the correlation function. Impulse observation control is useful in the infinite-dimensional filtering problems for distributed parameter systems, where it is necessary to urgently eliminate undesirable behavior of the correlation function (say, to terminate its fast transition to infinity) or to ensure fast high-accuracy estimation on a short time interval (if the noise in observations is too high).