Comparison of Three Kinds of Nonlinear Filter Methods

摘要

Extended Kalman filter is one of the most widely used methods for nonlinear system estimation. This paper introduces two new filtering algorithms, called unscented Kalman filtering (UKF) and particle filtering (PF). They can yield better performance than that of extended Kalman filtering (EKF) and have been shown to be a superior alternative to the EKF in a variety of applications, because UKF and PF do not involve the linearization approximating to nonlinear systems, that is required by the EKF. UKF uses a deterministic sampling approach. These sample points completely capture the true mean and covariance of the nonlinear system. The base idea of PF is the approximation of relevant probability distributions using the concepts of sequential importance sampling and approximation of probability distributions with a set of discrete random samples with associated weights. But these two methods still need to be improved in the aspects of accuracy and calculating speed.