Trellis based filtering
Robert J. Kee, G.W. Irwin · 2003
The work of K. Demirbas (1982, 1986, 1987, 1989) on using trellis diagrams for state estimation is extended. Techniques for improving the accuracy of the approach are considered, and a lack of robustness discovered in the original algorithm is highlighted. This is overcome by using the measurement equation to propagate the trellis, and the resulting filter is shown to be superior to the extended Kalman filter for some nonlinear examples.>