Paper Info
Title
Attitude Estimation Using Kalman Filtering: External Acceleration Compensation Considerations.
Abstract
Attitude estimation is often inaccurate during highly dynamic motion due to the external acceleration. This paper proposes extended Kalman filter-based attitude estimation using a new algorithm to overcome the external acceleration. This algorithm is based on an external acceleration compensation model to be used as a modifying parameter in adjusting the measurement noise covariance matrix of the extended Kalman filter. The experiment was conducted to verify the estimation accuracy, that is, one-axis and multiple axes sensor movement. Five approaches were used to test the estimation of the attitude: (1) the KF-based model without compensating for external acceleration, (2) the proposed KF-based model which employs the external acceleration compensation model, (3) the two-step KF using weighted-based switching approach, (4) the KF-based model which uses the threshold-based approach, and (5) the KF-based model which uses the threshold-based approach combined with a softened part approach. The proposed algorithm showed high effectiveness during the one-axis test. When the testing conditions employed multiple axes, the estimation accuracy increased using the proposed approach and exhibited external acceleration rejection at the right timing. The proposed algorithm has fewer parameters that need to be set at the expense of the sharpness of signal edge transition.
Year
DOI
Venue
2016
10.1155/2016/6943040
JOURNAL OF SENSORS
Field
DocType
Volume
Extended Kalman filter,Control theory,Kalman filter,Acceleration,Engineering,Covariance matrix,Signal edge
Journal
2016
ISSN
Citations 
PageRank 
1687-725X
2
0.45
References 
Authors
0
2
Name
Order
Citations
PageRank
Widodo, R.B.131.81
Chikamune Wada2219.55