On the solution of the linear estimation problem involving correlated signal and noise
Rosa M. Fernández-Alcalá, Jesús Navarro-Moreno, Juan Carlos Ruiz-Molina · 2003
An efficient solution is given to the problem of linear least-squares estimation of a continuous stochastic process corrupted by an additive white noise which is correlated with the signal. This solution is based on the approximate Karhunen-Loeve (KL) expansion and it can be derived through a recursive algorithm which is similar to a Kalman-Bucy filter.