La. Johnston et V. Krishnamurthy, Derivation of a sawtooth iterated extended Kalman smoother via the AECM algorithm, IEEE SIGNAL, 49(9), 2001, pp. 1899-1909
The iterated extended Kalman smoother (IEKS) is derived under expectation-m
aximization (EM) algorithm formalism, providing insight into the behavior o
f the suboptimal extended Kalman filter (EKF) and smoother (EKS). Through a
n investigation of smoothing algorithms that result from variants of the EM
algorithm, the sawtooth iterated extended Kalman smoother (SIEEKS) and its
computationally inexpensive counterparts are proposed via the alternating
expectation conditional maximization (AECM) algorithm. The SIEEKS is guaran
teed to produce a sequence estimate that moves up the likelihood surface. N
umerical simulations including frequency tracking examples display the supe
rior performance of the sawtooth EKF over the standard EKF for a range of n
onlinear signal models.