Kalman-filter control schemes for fringe tracking
Development and application to VLTI/GRAVITY
LESIA, Observatoire de Paris, CNRS, UPMC, Université Paris
Diderot, Paris Sciences et Lettres,
5 place Jules Janssen, 92195
2 Instituut voor Sterrenkunde, KU Leuven, Celestijnenlaan 200D, 3001 Leuven, Belgium
3 Groupement d’Intérêt Scientifique PHASE (Partenariat Haute résolution Angulaire Sol Espace) between ONERA, Observatoire de Paris, CNRS and Université Paris Diderot, France
Accepted: 28 February 2012
Context. The implementation of fringe tracking for optical interferometers is inevitable when optimal exploitation of the instrumental capacities is desired. Fringe tracking allows continuous fringe observation, considerably increasing the sensitivity of the interferometric system. In addition to the correction of atmospheric path-length differences, a decent control algorithm should correct for disturbances introduced by instrumental vibrations, and deal with other errors propagating in the optical trains.
Aims. In an effort to improve upon existing fringe-tracking control, especially with respect to vibrations, we attempt to construct control schemes based on Kalman filters. Kalman filtering is an optimal data processing algorithm for tracking and correcting a system on which observations are performed. As a direct application, control schemes are designed for GRAVITY, a future four-telescope near-infrared beam combiner for the Very Large Telescope Interferometer (VLTI).
Methods. We base our study on recent work in adaptive-optics control. The technique is to describe perturbations of fringe phases in terms of an a priori model. The model allows us to optimize the tracking of fringes, in that it is adapted to the prevailing perturbations. Since the model is of a parametric nature, a parameter identification needs to be included. Different possibilities exist to generalize to the four-telescope fringe tracking that is useful for GRAVITY.
Results. On the basis of a two-telescope Kalman-filtering control algorithm, a set of two properly working control algorithms for four-telescope fringe tracking is constructed. The control schemes are designed to take into account flux problems and low-signal baselines. First simulations of the fringe-tracking process indicate that the defined schemes meet the requirements for GRAVITY and allow us to distinguish in performance. In a future paper, we will compare the performances of classical fringe tracking to our Kalman-filter control.
Conclusions. Kalman-filter based control schemes will likely become the next standard for fringe tracking, providing the possibility to maximize the tracking performance. The results of the present study are currently being incorporated into the final design of GRAVITY.
Key words: techniques: interferometric / techniques: high angular resolution
© ESO, 2012