A Multiplatform CPU-Based Architecture for Cost-Effective Adaptive Optics Systems

Abstract

An Adaptive Optics (AO) system is composed of three key elements: a wave-front sensor that detects the aberrations; a deformable mirror that provides the wavefront correction and a closed-loop control system that elaborates the measurements acquired by the sensor and sends commands to the mirror. The control system can be implemented on a dedicated platform (e.g. FPGA) or on general-purpose platforms (e.g. CPU, GPU). Dedicated hardware guarantees high performance but needs more development time and programming skills than general-purpose hardware, leading to a less maintainable system for the end user. The proposed solution aims to be a cost-effective, multi-platform, CPU-based flexible framework. The software, developed in C++ and using Eigen and Qt libraries, provides the tools to tune and control the adaptive optics system, from wavefront measurement settings to controller parameters. A logging feature allows in-depth offline data analysis, while scripting enables execution of batch experiments. The adaptive optics system is tuned and evaluated by interfacing the wavefront sensor and deformable mirror with our software architecture. The results show that the proposed solution is able to correct the aberrations of a low to medium size SCAO system, with a control frequency up to 500Hz and computational latency of 0.04ms, using a consumer-grade notebook.