P. M. Chichura,A. Rahlin,A. J. Anderson,B. Ansarinejad,M. Archipley,L. Balkenhol,K. Benabed,A. N. Bender,B. A. Benson,F. Bianchini,L. E. Bleem,F. R. Bouchet,L. Bryant,E. Camphuis,J. E. Carlstrom,C. L. Chang,P. Chaubal,A. Chokshi,T.-L. Chou,A. Coerver,T. M. Crawford,C. Daley,T. de Haan,K. R. Dibert,M. A. Dobbs,M. Doohan,A. Doussot,D. Dutcher,W. Everett,C. Feng,K. R. Ferguson,K. Fichman,A. Foster,S. Galli,A. E. Gambrel,R. W. Gardner,F. Ge,N. Goeckner-Wald,R. Gualtieri,F. Guidi,S. Guns,N. W. Halverson,E. Hivon,G. P. Holder,W. L. Holzapfel,J. C. Hood,A. Hryciuk,N. Huang,F. Kéruzoré,A. R. Khalife,J. Kim,L. Knox,M. Korman,K. Kornoelje,C.-L. Kuo,K. Levy,A. E. Lowitz,C. Lu,A. Maniyar,D. P. Marrone,E. S. Martsen,F. Menanteau,M. Millea,J. Montgomery,Y. Nakato,T. Natoli,G. I. Noble,Y. Omori,S. Padin,Z. Pan,P. Paschos,K. A. Phadke,A. W. Pollak,K. Prabhu,W. Quan,M. Rahimi,C. L. Reichardt,M. Rouble,J. E. Ruhl,E. Schiappucci,J. A. Sobrin,A. A. Stark,J. Stephen,C. Tandoi,B. Thorne,C. Trendafilova,C. Umilta,J. Veitch-Michaelis,J. D. Vieira,A. Vitrier,Y. Wan,N. Whitehorn,W. L. K. Wu,M. R. Young,K. Zagorski,J. A. Zebrowski

Abstract:We present improvements to the pointing accuracy of the South Pole Telescope (SPT) using machine learning. The ability of the SPT to point accurately at the sky is limited by its structural imperfections, which are impacted by the extreme weather at the South Pole. Pointing accuracy is particularly important during SPT participation in observing campaigns with the Event Horizon Telescope (EHT), which requires stricter accuracy than typical observations with the SPT. We compile a training dataset of historical observations of astronomical sources made with the SPT-3G and EHT receivers on the SPT. We train two XGBoost models to learn a mapping from current weather conditions to two telescope drive control arguments -- one which corrects for errors in azimuth and the other for errors in elevation. Our trained models achieve root mean squared errors on withheld test data of $2.14''$ in cross-elevation and $3.57''$ in elevation, well below our goal of $5''$ along each axis. We deploy our models on the telescope control system and perform further in situ test observations during the EHT observing campaign in 2024 April. Our models result in significantly improved pointing accuracy: for sources within the range of input variables where the models are best trained, average combined pointing error improved 33%, from $15.9''$ to $10.6''$. These improvements, while significant, fall shy of our ultimate goal, but they serve as a proof of concept for the development of future models. Planned upgrades to the EHT receiver on the SPT will necessitate even stricter pointing accuracy which will be achievable with our methods.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to improve the pointing accuracy of the South Pole Telescope (SPT), especially in order to meet the extremely high - precision requirements for the Event Horizon Telescope (EHT) observation campaigns. ### Problem Background The SPT is located at the Amundsen - Scott South Pole Station in Antarctica and is mainly used to observe the cosmic microwave background radiation (CMB). However, due to extreme weather conditions (such as low temperature and large temperature differences), the structure of the SPT will be deformed, which will affect its pointing accuracy. Especially when participating in EHT observations, the SPT requires higher pointing accuracy because the observation targets of the EHT are celestial bodies such as black holes, which require extremely high angular resolution, and pointing errors cannot be corrected by post - processing. ### Solution To solve this problem, the author proposes to use a machine - learning (ML) model to correct the pointing error of the SPT in real - time. Specifically: 1. **Data Collection**: The author collected historical observation data, including the positions of astronomical sources, the current state of the telescope, and weather conditions. 2. **Model Training**: Two XGBoost models were trained to correct the azimuth and elevation errors respectively. These models take the current weather conditions as input and output the adjusted telescope drive control parameters. 3. **Model Deployment and Testing**: The trained models were integrated into the telescope control system and field - tested during the EHT observation campaign in April 2024. ### Results After testing, these ML models significantly improved the pointing accuracy of the SPT. For the observation sources within the model training range, the average combined pointing error was reduced from 15.9" to 10.6", an improvement of 33%. Although this improvement has not yet fully reached the expected goal, it provides a strong proof - of - concept for future further optimization. ### Formula Summary The main formulas involved in the paper are as follows: - **Gravity Deflection Correction**: \[ \delta_1=\delta_0 - \phi_0\sin\delta_0 - \phi_1\cos\delta_0 \] \[ \alpha_1 = \alpha_0 \] - **Azimuth - Axis Tilt Correction**: \[ \Omega=- 180^{\circ}-\tan^{-1}\left(\frac{\sin(\tau_{ax}\cos\tau_{ay})}{\cos(\tau_{ax}\cos\tau_{ay})\sin\tau_{ay}}\right) \] \[ \rho=\cos^{-1}(\cos\tau_{ay}\cos(\tau_{ax}\cos\tau_{ay})) \] \[ \delta_2=\sin^{-1}(\sin\delta_1\cos\rho - \cos\delta_1\sin\rho\sin\omega) \] \[ \alpha_2=\Omega-\tan^{-1}\left(\frac{\cos\omega\cos\delta_2}{-\cos\rho\sin\omega\cos\delta_2 - \sin\rho\sin\delta_2}\right) \] where $\omega\equiv\Omega - 90^{\circ}-\alpha_1$. - **Other Correction Steps**: Including elevation - axis tilt correction, optical system alignment error correction, and encoder offset correction. See the original paper for specific formulas. Through these methods, the author shows how to use machine - learning technology to significantly improve the pointing accuracy of the South Pole Telescope, so as to better support high - precision astronomical observations.

Pointing Accuracy Improvements for the South Pole Telescope with Machine Learning

HIGH-PRECISION ASTROMETRY WITH A DIFFRACTIVE PUPIL TELESCOPE

Machine Learning based Pointing Models for Radio/Sub-millimeter Telescopes

Performance and characterization of the SPT-3G digital frequency-domain multiplexed readout system using an improved noise and crosstalk model

The Design and Integrated Performance of SPT-3G

The 10 Meter South Pole Telescope

A VLBI receiving system for the South Pole Telescope

The First Public Release of South Pole Telescope Data: Maps of a 95-square-degree Field from 2008 Observations

Testing the $\mathbfΛ$CDM Cosmological Model with Forthcoming Measurements of the Cosmic Microwave Background with SPT-3G

Year Two Instrument Status Of The Spt-3g Cosmic Microwave Background Receiver

Design and Bolometer Characterization of the SPT-3G First-year Focal Plane

Prototype System for Unmanned Reference Point Determination at the Sub‐Millimeter Level

Monitoring the pointing of the prototype LST-1 using star reconstruction in the Cherenkov camera

Measuring Track-Related Pointing Errors on the Nanshan Radio Telescope with an Optical Pointing Telescope

Measurement and Modeling of Polarized Atmosphere at the South Pole with SPT-3G

Detection of Thermal Emission at Millimeter Wavelengths from Low-Earth Orbit Satellites

Correcting Turbulence-induced Errors in Fiber Positioning for the Dark Energy Spectroscopic Instrument

Telescope performance real-time monitoring based on machine learning

Estimation of Pointing Errors of Large Radio Telescopes under Solar Radiation Based on Digital Twin

SPT clusters with DES and HST weak lensing. I. Cluster lensing and Bayesian population modeling of multiwavelength cluster datasets