Aircraft Inertial Measurement Unit Fault Diagnosis Based on Optimal Two-Stage UKF

基于OTSUKF的飞行器惯性测量单元的故障诊断

¹ School of Automation, Northwestern Polytechnical University, Xi’an 710129, China
² Aviation Ammunition Research Institute of Weapon Industry Group, Harbin 150036, China

Received: 9 September 2017

Abstract

Optimal two stage Kalman filter (OTSKF) is able to obtain optimal estimation of system states and bias for linear system which contains random bias. Unscented Kalman filter (UKF) is a conventional nonlinear filtering method which utilizes Sigmas point sampling and unscented transformation technology realizes propagation of state means and covariances through nonlinear system. Aircraft is a typical complicate nonlinear system, this paper treats the faults of Inertial Measurement Unit (IMU) as random bias, established a filtering model which contains faults of IMU. Hybird the two stage filtering technique and UKF, this paper proposed an optimal two stage unscented Kalman filter (OTSUKF) algorithm which is suitable for fault diagnosis of IMU, realized optimal estimation of system states and faults identification of IMU via proposed innovative designing method of filtering model and the algorithm was validated that it is robust to wind disterbance via real flight data and it is also validated that proposed OTSUKF is optimal in the existance of wind disturbance via comparing with the existance iterated optimal two stage extended kalman filter (IOTSEKF) method.

摘要

线性系统存在随机偏差情况下，最优二步卡尔曼滤波（OTSKF）可以获得系统状态及偏差的最优估计。无迹卡尔曼滤波（UKF）利用Sigma点采样和UT变换技术经非线性系统传播状态的均值和方差，是一种典型的非线性滤波方法。飞行器是一种典型复杂非线性系统，将惯性测量单元（IMU）的故障作为一种随机偏差处理，建立了包含IMU故障的滤波模型。将UKF算法和二步滤波思想应用到飞行器之中，提出了一种适用于飞行器IMU故障诊断的最优二步无迹卡尔曼滤波（OTSUKF）算法。针对于飞行器，提出了一种滤波模型设计的方法。基于该模型，应用所提出的OTSUKF算法实现了飞行器状态的最优估计和IMU故障的辨识，该算法经过了实际飞行数据验证其对风扰动具有鲁棒性并且与已经被提出的迭代最优二步扩展卡尔曼滤波（IOTSEKF）方法进行了对比验证其最优性。