The Unscented Kalman Filter for Nonlinear Estimation无迹卡尔曼非线性滤波方法.pdfVIP

The Unscented Kalman Filter for Nonlinear Estimation Eric A. Wan and Rudolph van der Merwe Oregon Graduate Institute of Science Technology 20000 NW Walker Rd, Beaverton, Oregon 97006 ericwan@, rvdmerwe@ Abstract 1. Introduction The Extended Kalman Filter (EKF) has become a standard The EKF has been applied extensively to the field of non- technique used in a number of nonlinear estimation and ma- linear estimation. General application areas may be divided chine learning applications. These include estimating the into state-estimation and machine learning. We further di- state of a nonlinear dynamic system, estimating parame- vide machine learning into parameter estimation and dual ters for nonlinear system identification (e.g., learning the estimation. The framework for these areas are briefly re- weights of a neural network), and dual estimation (e.g., the viewed next. Expectation Maximization (EM) algorithm) where both states State-estimation and parameters are estimated simultaneously. This paper points out the flaws in using the EKF, and The basic framework for the EKF involves estimation of the introduces an improvement, the Unscented Kalman Filter state of a discrete-time nonlinear dynamic system, (UKF), proposed by Julier and Uhlman [5]. A central and (1) vital operation performed in the Kalman Filter is the prop- (2)