A NON CONVENTIONAL UAV IN GROUND EFFECT: ESTIMATION OF STATE AND TURBULENCE VIA EXTENDED KALMAN FILTER

Caterina Grillo, Cinzia Gatto, Francesco Paolo Vitrano

Risultato della ricerca: Articlepeer review

Abstract

ABSTRACTThis paper discusses the synthesis of an Extended Kalman Filter (EKF) to perform both wind velocities and stateestimation for a non conventional UAV flying in ground effect. Since, in IGE flight, motions into the symmetry plane areof primary concern, the study focuses on the longitudinal aircraft dynamics. The proposed estimator requiresmeasurement of few flight variables, easily obtainable by means of conventional sensors; besides, it does not useInertial Measurement Unit (IMU). To simulate a low cost sensing equipment, the model outputs are corrupted by whitenoise of relatively high standard deviation. Furthermore, to cope with the low rate of the GPS with respect to the othersensors, the EKF algorithm is modified to allow for a dual rate measurement model. State propagation is obtained bymeans of an accurate, highly non-linear mathematical model of the UAV dynamics which allows to cope with both thenon conventional configuration and ground effect. Combined estimation of the UAV state and wind velocity is obtainedby augmenting the system state vector with the wind velocity components. Numerical simulations show the effectivenessof the proposed method in widely different maneuvering conditions mixing IGE and OGE flight stages and in presenceof largely dissimilar wind disturbance characteristics. The performance advantages with respect to a classical,computationally simpler estimation scheme are also highlighted. Closed loop simulations have shown the suitability ofthe obtained state estimates for control purposes.
Lingua originaleEnglish
pagine (da-a)1-11
Numero di pagine11
RivistaAUTOMATIC CONTROL IN AEROSPACE
Volume1
Stato di pubblicazionePublished - 2009

Fingerprint

Entra nei temi di ricerca di 'A NON CONVENTIONAL UAV IN GROUND EFFECT: ESTIMATION OF STATE AND TURBULENCE VIA EXTENDED KALMAN FILTER'. Insieme formano una fingerprint unica.

Cita questo