Investigation of Adaptive Robust Kalman Filtering Algorithms for GPS/DR Navigation System Filters

2016 2nd International Conference on Mechanical and Aeronautical Engineering (ICMAE 2016)

机械制造;航空航天工程

Elzoghby, Mostafa^1 ; Arif, Usman^1 ; Li, F.U.^1 ; Zhi Yu, X.I.^1

School of Automation Science and Electrical Engineering, Beihang University, Beijing

100191, China^1

关键词: Aerial vehicle;    Kalman filter algorithms;    Noise covariance;    Real-time application;    Robust Kalman filtering;    Robust Kalman filters;    Sage-husa;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/187/1/012019/pdf
DOI  :  10.1088/1757-899X/187/1/012019

学科分类：航空航天科学

来源: IOP

【 摘 要 】

The conventional Kalman filter (KF) algorithm is suitable if the characteristic noise covariance for states as well as measurements is readily known but in most cases these are unknown. Similarly robustness is required instead of smoothing if states are changing abruptly. Such an adaptive as well as robust Kalman filter is vital for many real time applications, like target tracking and navigating aerial vehicles. A number of adaptive as well as robust Kalman filtering methods are available in the literature. In order to investigate the performance of some of these methods, we have selected three different Kalman filters, namely Sage Husa KF, Modified Adaptive Robust KF and Adaptively Robust KF, which are easily simulate able as well as implementable for real time applications. These methods are simulated for land based vehicle and the results are compared with conventional Kalman filter. Results show that the Modified Adaptive Robust KF is best amongst the selected methods and can be used for Navigation applications.