The Ukf Algorithm And Improved Algorithm

The Kalman Filter (KF) is one of the most widely used methods for tracking and estimation due to its simplicity, optimality, tractability and robustness. However, the application of the KF to nonlinear systems can be quite difficult. Using the principle that a set of discretely sampled points can be used to parameterize mean and covariance, the Unscented Kalman Filter (UKF) yields performance equivalent to the KF for linear systems yet generalizes elegantly to nonlinear systems without the linearization steps required by the EKF (Extended Kalman Filter). Along with the rapid development of UKF in the non-linear filtering, a growing number of scientists begin to realize its great potential to handle non-linear propagation problem.This paper carefully investigates the performance, weaknesses and improvement in UKF algorithm.First of all, this paper introduces and analyzes three types of sampling strategy. Moreover, a new adaptive sampling algorithm is introduced and tested in this paper. Then, a detailed discussion about the algorithm is made in this paper. And so will the range of its application.Secondly, this paper examines UKF algorithms which are used to compensate the lack of the priori knowledge of the process uncertainty distribution and to improve the performance of Kalman Filter for the state and parameter estimations. In order to diminish the repercussion of linearization error on estimation of state of nonlinear system, this paper adopts Fading Memory algorithm, a division of Kalman Filter, to implement the filtering of nonlinear system with high level of accuracy. On the other hand, the aim of the researches on the adaptive fading and strong tracking strategy is to enhance the robustness to neutralize errors brought by the inaccurate system models and the capability of tracking the abrupt state.Finally, to improve the estimation accuracy and the convergence speed of the UKF, a novel adaptive filter method is adopted in this paper. This method can only be used in hybrid propagation function. On the basis of covariance matching method and scaling of process noise, the designated adaptive mechanism drives the filter autonomously to the optimal mode.