Filtrage et lissage statistiques optimaux linéaires / J.-C. Radix

This course introduces linear filtering and smoothing from an engineering viewpoint. The mathematical knowledge in probability theory and linear algebra which is needed is elementary and is reviewed; on the other hand, several examples are worked out; the equations are clearly written and intuitive explanation of the algorithms and results is emphasized. Firstly, Kalman filtering is studied, both is continuous and discrete time; the various classical generalizations of the initial problem are also discussed (coloured noise, correlated noise, nonlinear systems). Secondly, the different classical formulations of the smoothing problem are stated and solved; examples of algorithms are provided with graphical representations of the results. Finally, after a theoretical study of the factorization of matrices, square root filtering algorithms are derived; in an appendix computer programs are given. (Zentralblatt)

Bibliogr. p. 240-243 et 350