Laetitia TOSONI defended her PhD on jan. 10th, 2025.
Place : amphithéâtre Emilie Châtelet, la bibliothèque de l’INSA Lyon. 69100 Villeurbanne
Jury :
Rapporteurs :
Mme. Christine PRELLE, Professeure, Université de technologie de Compiègne
M. Sébastien CAUET , Professeur, Université de Poitiers
Examinateurs :
M. Stéphane CARO, Directeur de Recherche, CNRS - LS2N
Mme. Caroline KULCSÁR, Professeure, Institut d’Optique Graduate School
Encadrement :
M. Minh Tu PHAM, Professeur, INSA Lyon
M. Paolo MASSIONI, Maitre de conférences HdR, INSA Lyon
Invité:
M. Elliot BROUSSARD, Ingénieur, Thales.
Abstract :
Modern communication needs require both an important data rate and a large geographic coverage. Satellite communication (Satcom) fulfills both these requirements, but demands that the antenna be pointed towards the target satellite for the entire duration of the communication. In the case of a mobile application, called SOTM (Satcom On-The-Move), the antenna is installed on a mechatronic system that maintaining an adequate pointing despite the carrier vehicle’s motions.
This mechatronic system must be controlled, and therefore a position setpoint must be computed. Several methods exist to compute this setpoint. Some exploit a signal sent by the satellite, but those are dependant on the availability of this signal. Others use geometric considerations based on inertial information provided by the SOTM station’s on-board sensors, but some unknown factors in the system’s geometry degrade the result’s precision beyond what is required for antenna pointing.
These unknown factors can be bypassed if the carrier vehicle is equipped with an inertial navigation system : the issue then comes back to estimating the spatial rotation between two inertial platforms ; one located on the carrier vehicle, the other located on the antenna. Transfer alignment is an effective method to estimate a spatial rotation between two inertial platforms, but it is traditionally applied to small and stationary spatial rotations. Estimation of large and time-varying spatial rotations use near-exclusively nonlinear models.
This PhD contributes two transfer alignment algorithms for arbitrary, time-varying spatial rotations. The first one, based on a measurement model from the existing literature, uses linear velocity and attitude measurements, a classic combination in the domain of transfer alignment. The second one uses linear acceleration and rotation rate measurements, rarely used in this domain but more suited to the hardware limitations of a SOTM application. An original propagation model, applicable to both cases, is also presented. Quaternions were chosen for attitude representation in order to keep the models linear.
The proposed algorithms were tested in simulation and experimented on a test bench. Appropriate metrics were defined in order to analyze algorithm performance and to identify error factors affecting it.
Keywords:
Transfer alignment, Real-time estimator, Satcom On The Move, Inertial stabilization, Kalman filtering