Details
| Originalsprache | Englisch |
|---|---|
| Aufsatznummer | 064041 |
| Seitenumfang | 11 |
| Fachzeitschrift | Physical review applied |
| Jahrgang | 20 |
| Ausgabenummer | 6 |
| Publikationsstatus | Veröffentlicht - 22 Dez. 2023 |
Abstract
Suspended optics in gravitational-wave (GW) observatories are susceptible to alignment perturbations, particularly slow drifts over time, due to variations in temperature and seismic levels. Such misalignments affect the coupling of the incident laser beam into the optical cavities, degrade both the circulating power and optomechanical photon squeezing, and thus decrease the astrophysical sensitivity to merging binaries. Traditional alignment techniques involve differential wave-front sensing using multiple quadrant photodiodes but are often bandwidth restricted and limited by the sensing noise. We present a successful implementation of neural-network-based sensing and control at a GW observatory and demonstrate low-frequency control of the signal-recycling mirror at the GEO 600 detector. Alignment information for three critical optics is simultaneously extracted from the interferometric dark-port camera images via a convolutional neural net-long short-term memory network architecture and is then used for multiple-input-multiple-output control using soft actor-critic-based deep reinforcement learning. The overall sensitivity improvement achieved using our scheme demonstrates the capabilities of deep learning as a viable tool for real-time sensing and control for current and next-generation GW interferometers.
ASJC Scopus Sachgebiete
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
in: Physical review applied, Jahrgang 20, Nr. 6, 064041, 22.12.2023.
Publikation: Beitrag in Fachzeitschrift › Artikel › Forschung › Peer-Review
}
TY - JOUR
T1 - Neural sensing and control in a kilometer-scale gravitational-wave observatory
AU - Mukund, N.
AU - Lough, J.
AU - Bisht, A.
AU - Wittel, H.
AU - Nadji, S.
AU - Affeldt, C.
AU - Bergamin, F.
AU - Brinkmann, M.
AU - Kringel, V.
AU - Lück, H.
AU - Weinert, M.
AU - Danzmann, K.
N1 - Funding Information: N.M. thanks Thomas Künzel, Franziska Albers, and MathWorks for their technical support. N.M. thanks Kong Chun-Wei, University of Michigan, for his insights about training deep RL controllers. We thank Lisa Barsotti and the LIGO control systems working group for their valuable suggestions. We are grateful for the computational resources provided by the LIGO Laboratory and supported by National Science Foundation Grants No. PHY-0757058 and No. PHY-0823459. We thank the GEO collaboration for the development and construction of GEO 600. We also extend our thanks to Walter Graß for his work keeping the interferometer in a good running state. We are grateful for support from the Science and Technology Facilities Council (STFC), the University of Glasgow in the United Kingdom, the Max Planck Society, the Bundesministerium für Bildung und Forschung (BMBF), the Volkswagen Stiftung, the cluster of excellence QUEST (Centre for Quantum Engineering and Space-Time Research), the International Max Planck Research School (IMPRS), and the State of Niedersachsen in Germany.
PY - 2023/12/22
Y1 - 2023/12/22
N2 - Suspended optics in gravitational-wave (GW) observatories are susceptible to alignment perturbations, particularly slow drifts over time, due to variations in temperature and seismic levels. Such misalignments affect the coupling of the incident laser beam into the optical cavities, degrade both the circulating power and optomechanical photon squeezing, and thus decrease the astrophysical sensitivity to merging binaries. Traditional alignment techniques involve differential wave-front sensing using multiple quadrant photodiodes but are often bandwidth restricted and limited by the sensing noise. We present a successful implementation of neural-network-based sensing and control at a GW observatory and demonstrate low-frequency control of the signal-recycling mirror at the GEO 600 detector. Alignment information for three critical optics is simultaneously extracted from the interferometric dark-port camera images via a convolutional neural net-long short-term memory network architecture and is then used for multiple-input-multiple-output control using soft actor-critic-based deep reinforcement learning. The overall sensitivity improvement achieved using our scheme demonstrates the capabilities of deep learning as a viable tool for real-time sensing and control for current and next-generation GW interferometers.
AB - Suspended optics in gravitational-wave (GW) observatories are susceptible to alignment perturbations, particularly slow drifts over time, due to variations in temperature and seismic levels. Such misalignments affect the coupling of the incident laser beam into the optical cavities, degrade both the circulating power and optomechanical photon squeezing, and thus decrease the astrophysical sensitivity to merging binaries. Traditional alignment techniques involve differential wave-front sensing using multiple quadrant photodiodes but are often bandwidth restricted and limited by the sensing noise. We present a successful implementation of neural-network-based sensing and control at a GW observatory and demonstrate low-frequency control of the signal-recycling mirror at the GEO 600 detector. Alignment information for three critical optics is simultaneously extracted from the interferometric dark-port camera images via a convolutional neural net-long short-term memory network architecture and is then used for multiple-input-multiple-output control using soft actor-critic-based deep reinforcement learning. The overall sensitivity improvement achieved using our scheme demonstrates the capabilities of deep learning as a viable tool for real-time sensing and control for current and next-generation GW interferometers.
UR - http://www.scopus.com/inward/record.url?scp=85180946467&partnerID=8YFLogxK
U2 - 10.1103/PhysRevApplied.20.064041
DO - 10.1103/PhysRevApplied.20.064041
M3 - Article
AN - SCOPUS:85180946467
VL - 20
JO - Physical review applied
JF - Physical review applied
SN - 2331-7019
IS - 6
M1 - 064041
ER -