State and force observers based on multibody models and the indirect Kalman filter. (June 2018)

• Record Type: Journal Article
• Authors: Sanjurjo, Emilio; Dopico, Daniel; Luaces, Alberto; Naya, Miguel Ángel
• Is Part Of: Mechanical systems and signal processing. Volume 106(2018)
• Page Start: 210
• Page End: 228
• Publication Date: 2018-06
• Subjects: Multibody dynamics -- Kalman filter -- State observer -- Force estimation; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2017.12.041
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: Presentation of three indirect Kalman filters based on multibody models. General expressions for the Jacobians of the plant and measurement models. Presentation of a new method providing state and input (force) estimation. Comparison and benchmarking of the three methods. Abstract: The aim of this work is to present two new methods to provide state observers by combining multibody simulations with indirect extended Kalman filters. One of the methods presented provides also input force estimation. The observers have been applied to two mechanism with four different sensor configurations, and compared to other multibody-based observers found in the literature to evaluate their behavior, namely, the unscented Kalman filter (UKF), and the indirect extended Kalman filter with simplified Jacobians (errorEKF). The new methods have some more computational cost than the errorEKF, but still much less than the UKF. Regarding their accuracy, both are better than the errorEKF. The method with input force estimation outperforms also the UKF, while the method without force estimation achieves results almost identical to those of the UKF. All the methods have been implemented as a reusable MATLAB® toolkit which has been released as Open Source inhttps://github.com/MBDS/mbde-matlab .