SUKF (Self-calibration Unscented Kalman Filter)

Abstract

The present invention provides a self-calibration unscented Kalman filter method, the steps are as follows: one: establish a nonlinear discrete system including unknown input; two: initialize the system composed of formula (1) and formula (2); three: Time update the system; four: combined with the measurement information, measure and update the state one-step prediction and one-step prediction error variance matrix; five: perform iterative calculation; through the above steps, the present invention introduces the self-calibration technology into the unscented Kalman Filtering method, the newly obtained self-calibrating unscented Kalman filtering method firstly solves the modeling problem of the nonlinear system affected by the unknown input. The adverse effect of filtering results greatly reduces the phenomenon of filtering divergence, thereby improving the filtering accuracy, and at the same time, as a filtering method that can resist system uncertainty, it improves the robustness of the system.

Description

【Technical field】 [0001] The invention provides a self-calibration unscented Kalman filtering method, and belongs to the technical field of nonlinear robust filtering. 【Background technique】 [0002] Since Kalman proposed the Kalman filtering method for linear systems in 1960, this method has been widely and deeply applied in various engineering fields. Based on the standard Kalman filter method, researchers have successively developed a series of methods such as Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), and Particle Filter (PF). It is used to solve the filtering problem of nonlinear systems. However, the extended Kalman filter method can only deal with weakly nonlinear systems. Although the particle filter method can deal with strongly nonlinear systems, there are problems of particle degradation and particle impoverishment. The most important method for the state estimation problem of linear systems. [0003] The unscented Kalman filter method is esta...

AI Technical Summary

Problems solved by technology

However, this method is only suitable for linear systems, and does not solve the problem that the state equation is affected by unknown inputs in nonlinear systems that are common in engineering practice.

Images