real-time adaptive optics (AO) control and low-latency edge computing
Abstract
Large ground-based astronomical telescopes deliver blurry images due to perturbations introduced by Earth's atmosphere. To recover sharp images, these distortions must be optically corrected in real time by an adaptive optics (AO) system. The control system that performs this task must operate thousands of optical sensors and correction elements at kHz speed to keep up with the fast-changing atmospheric turbulence. We are developing a free and open-source software (FOSS) package for high-performance, low-latency computing pipelines with stringent real-time requirements for astronomical adaptive optics (AO). The Compute and Control for Adaptive Optics (CACAO) toolkit’s modular design allows for customization for multiple applications, including imaging life-bearing planets orbiting nearby stars, which requires exquisite image quality with large telescopes. Over the last year, the software framework has undergone major improvements for modularity, ease-of-use, and support for modern AI/ML techniques.