Authors

Anne Costille (1), Thierry Fusco (2)

Affiliations

(1) IPAG, (2) ONERA

Abstract

In the frame of the design of new instruments for the Extremely Large Telescopes (ELT), new techniques of Adaptive Optics have been developed. These new techniques, generically called Wide Field Adaptive Optics (WFAO), are based on a tomographic reconstruction of the turbulent volume followed by a projection onto DM(s) in order to ensure a good correction either in dedicated directions (LTAO or MOAO schemes) or in a large Field of View (GLAO or MCAO schemes). All these systems require a 3D phase reconstruction and thus a statistical representation of the turbulent volume through the knowledge of the Cn² profile. It matters both for an accurate simulation of the input perturbations in the case of performance analysis and system design, but also for an efficient model description in the tomographic reconstruction process. Firstly, we discuss and analyze the impact of the structure and the complexity of the real Cn² profile onto the WFAO performance. We demonstrate that a classical integrated parameter such as θ_0 is not sufficient and that a more complex criterion including information on Cn² distribution is mandatory. Then, we focus on the impact of Cn² model error in the tomographic reconstruction process with respect to the input profile. We demonstrate that number and position of layers are two critical parameters but we also highlight the fact that in a regularized (MMSE-like) scheme, increasing the number of layers in the model will relax the requirements on the knowledge of their absolute positions. In conclusion, we demonstrate that it is critical to have access to high resolution Cn² profile with few hundred meters resolution to ensure a good performance evaluation of a WFAO system. We also give a first quantification of the number and accuracy of the prior information on Cn² to have an efficient tomographic reconstruction process.