Adaptive Observer and Kalman Filtering

Khadidja Benzemrane (1) , Gilney Damm (1) , Giovanni L. Santosuosso (2)

1 IBISC - Informatique, Biologie Intégrative et Systèmes Complexes
2 Dipartimento di Ingegneria Elettronica

In this paper the problem of the speed estimation of an Unmanned Aerial Vehicle is addressed, when only the standard outputs (acceleration, angles and angular speeds) are available for measurement. We focus our analysis on a prototype drone - a 4 rotors helicopter robot- which is not equipped with GPS related devices and relies on the Inertial Measurement Unit (IMU) only. Two different approaches have been compared. The first one uses a classical method based on Kalman Filtering while the second solution is provided in the frame of adaptive observation theory. These estimators have been tested in two situations : when exact measurements are available and in the more realistic case of noisy acceleration measurements.

Mots clés: Robust estimation; Nonlinear observer and filter design

Domaines: Automatique / Robotique

Khadidja Benzemrane, Gilney Damm, Giovanni L. Santosuosso. Adaptive Observer and Kalman Filtering. 17th World Congress, International Federation of Automatic Control, IFAC, Jul 2008, Seoul, South Korea. pp.3865--3870, ⟨10.3182/20080706-5-KR-1001.3675⟩. ⟨hal-00342773⟩