The Real Time Analysis framework of the Cherenkov Telescope Array's Large-Sized Telescope

Sami Caroff, Kazuki Abe, Shotaro Abe, Arnau Aguasca-Cabot, I. Agudo, Nuria Álvarez Crespo, L. A. Antonelli, C. Aramo, Axel Arbet-Engels, Cornelia Arcaro, Manuel Artero, Katsuaki Asano, Ali Baktash, Aya Bamba, Andrés Baquero, Leonardo Baroncelli, U. Barres de Almeida, J. A. Barrio, Ivana Batković, Joshua Ryo Baxter, Josefa Becerra González, E. Bernardini, Mabel Bernardos, Juan Bernete Medrano, A. Berti, Pooja Bhattacharjee, Noah Biederbeck, C. Bigongiari, E. Bissaldi, O. Blanch, G. Bonnoli, P. Bordas, A. Bulgarelli, Irene Burelli, L. Burmistrov, Mario Buscemi, Martina Cardillo, A. Carosi, Marie-Sophie Carrasco, F. Cassol-Brunner, D. Cauz, Davide Cerasole, Giovanni Ceribella, Yating Chai, Ks Cheng, A. Chiavassa, Michiyuki Chikawa, L. Chytka, Alejo Cifuentes, J. L. Contreras, J. Cortina, H. Costantini, M. Dalchenko, F. Dazzi, Alessandro De Angelis, Mathieu de Bony de Lavergne, B. De Lotto, Mario de Lucia, Raniere de Menezes, L. del Peral, Guillaume Deleglise, C. Delgado Mendez, J. Delgado, Domenico Della Volpe, Michael Dellaiera, Ambra Di Piano, F. Di Pierro, A. Di Pilato, Riccardo Di Tria, L. Di Venere, C. Díaz, Rune Michael Dominik, D. Dominis Prester, Alice Donini, D. Dorner, M. Doro, Laura Eisenberger, Dominik Elsässer, Gabriel Emery, Juan Escudero, V. Fallah Ramazanı, Giovanna Ferrara, F. Ferrarotto, A. Fiaßon, L. Foffano, Lluís Freixas Coromina, Stefan Fröse, Satoshi Fukami, Yasushi Fukazawa, Enrique Rivera Garcí­a, R. J. Garcı́a López, Claudio Gasbarra, Dario Gasparrini, Felix Geyer, Joao Giesbrecht Paiva, N. Giglietto, F. Giordano, Paweł Gliwny, N. Godinović, Roger Grau Haro2023 · DOI

TL;DR

This paper introduces the Real-Time Analysis (RTA) framework for the Cherenkov Telescope Array's (CTA) Large-Sized Telescope (LST-1). The RTA is designed for rapid processing of gamma-ray data, enabling quick identification of transient events like gamma-ray bursts, crucial for follow-up observations and understanding the dynamic universe.

What Problem Does This Paper Address?

Gamma-ray astronomy aims to study high-energy cosmic phenomena, including transient events like gamma-ray bursts (GRBs) and flaring active galactic nuclei (AGN). The detection of these events at very high energies (VHE, > 100 GeV) has opened new research avenues. The Cherenkov Telescope Array (CTA) is a next-generation observatory designed to significantly advance this field. A key challenge for studying transient phenomena is the need for rapid detection and analysis. This allows astronomers to quickly identify these rare and fleeting events and initiate follow-up observations. The LST-1, a prototype for CTA, is specifically designed with a low energy threshold to detect these phenomena. To fully leverage this capability, a robust and fast real-time analysis system is essential to process data as it's collected, providing timely alerts and feedback to operators.

Methods

The RTA framework is built upon a multi-stage data processing pipeline. The core of the analysis involves:

  • R0->DL1 Stage: This initial stage handles data calibration, image aggregation, and cleaning. It extracts fundamental Hillas and timing parameters from the raw telescope data. This stage is crucial for reducing the massive raw data volume to a manageable size.
  • DL1->DL2 Stage: This stage focuses on evaluating key event parameters: energy, direction, and a "gammaness" score, which classifies events as likely gamma rays or background.
  • DL2->DL3 Stage: The final stage involves selecting gamma-like events, further reducing the data flow to a very small size, suitable for storage and subsequent high-level analysis.

The software architecture is designed for efficiency, utilizing C++ for computationally intensive tasks and leveraging techniques like vectorization to optimize performance. A workload manager (Slurm) is employed for resource and process management across dedicated servers. The RTA's offline analysis chain is used for high-level parameter computation to simplify maintenance, with a conversion step to align with the CTAO standard DL1 format.

Datasets

The paper describes the use of:

  • Real-time data from LST-1: This data is collected during actual observations.
  • Monte Carlo simulations: These simulations are used to train random forests for event classification (gammaness) and to evaluate performance metrics like effective area, energy bias, and resolution. The simulations are generated for gamma-ray events with a specific zenith angle (20°).
  • Markarian 421 data: This dataset, consisting of 14 observations totaling 4.2 hours, was used to test and validate the reconstruction performance of the RTA against the established offline analysis pipeline (lstchain).

Key Insights and Ideas

  • Real-time analysis is critical for transient astronomy: The paper strongly emphasizes that the ability to analyze data in real-time is not just a convenience but a necessity for capturing and studying transient gamma-ray sources.
  • Optimization is key for speed: The RTA framework employs several optimization techniques, including C++ implementation, vectorization, and efficient data reduction stages, to achieve the required processing speed.
  • RTA performance is comparable to offline analysis: While there are some differences, the RTA demonstrates performance metrics that are generally competitive with the established offline analysis chain, particularly for rapid detection use cases.
  • Scalability for the full CTA: The software architecture is designed with scalability in mind, aiming to accommodate the much larger data volumes expected from the full CTA observatory.

Key Results

  • Processing Speed: The RTA can produce DL3 files from 20,000 R0 events in approximately 66 seconds. The R0->DL1 stage achieves a computing speed of 0.5 ms/event, significantly faster than the offline chain's 0.03 s/event.
  • Event Rate Handling: The RTA was tested and found to be unaffected by event rate peaks of up to 14 kHz, demonstrating its ability to handle high data throughput.
  • Reconstruction Performance: When tested on Markarian 421 data, the RTA achieved a maximum significance of 15.3σ with a gammaness threshold of 0.75. While the offline analysis achieved a higher significance (31.2σ with 0.7 gammaness), the RTA's performance is considered promising for real-time applications.
  • Sensitivity: The RTA's differential sensitivity is found to be roughly two times worse than the offline analysis chain above 0.4 TeV, and worsens at lower energies. However, for detecting transient events in shorter time scales (e.g., 20 minutes), the RTA shows promise, capable of detecting an integrated flux of 0.3 Crab units in a single run.

Limitations and Future Work

  • Differential Sensitivity: The RTA's differential sensitivity is noted as being worse than the offline analysis chain, particularly at higher energies. This is attributed to factors like a simplified linear regression for timing parameter computation and a more basic calibration process compared to the full offline chain.
  • Angular Resolution: The RTA's angular resolution is reported to be worse than the lstchain, contributing to the difference in effective area.
  • Future Work: The paper highlights that the RTA will be integrated into the Array Control and Data Acquisition System (ACADA) for CTAO's first release. Future work will involve adapting the software for the full array, including modifications to DL1 for merging triggered events and DL2 for stereoscopic reconstruction. The scalability and modularity of the current design are intended to facilitate these future developments.

Summary updated

Abstract

The Large-Sized Telescopes (LSTs) of the Cherenkov Telescope Array Observatory (CTAO) will play a crucial role in the study of transient gamma-ray sources, such as gamma-ray bursts and flaring active galactic nuclei. The low energy threshold of LSTs makes them particularly well suited for the detection of these phenomena. The ability to detect and analyze gamma-ray transients in real-time is essential for quickly identifying and studying these rare and fleeting events. In this conference, we will present recent advances in the real-time analysis of data from the LST-1, the first prototype of LST located in the Canary island of La Palma. We will discuss in particular the development of new algorithms for event reconstruction and background rejection. These advances will enable rapid identification and follow-up observation of transient gamma-ray sources, making the LST-1 a powerful tool for the study of the dynamic universe. The implementation of this framework in the future Array Control and Data Acquisition System (ACADA) of CTAO will be discussed as well, based on the experience with LST.
BibTeX
@misc{caroff2023real,
  author = {Sami Caroff and Kazuki Abe and Shotaro Abe and Arnau Aguasca-Cabot and I. Agudo and Nuria Álvarez Crespo and L. A. Antonelli and C. Aramo and Axel Arbet-Engels and Cornelia Arcaro and Manuel Artero and Katsuaki Asano and Ali Baktash and Aya Bamba and Andrés Baquero and Leonardo Baroncelli and U. Barres de Almeida and J. A. Barrio and Ivana Batković and Joshua Ryo Baxter and Josefa Becerra González and E. Bernardini and Mabel Bernardos and Juan Bernete Medrano and A. Berti and Pooja Bhattacharjee and Noah Biederbeck and C. Bigongiari and E. Bissaldi and O. Blanch and G. Bonnoli and P. Bordas and A. Bulgarelli and Irene Burelli and L. Burmistrov and Mario Buscemi and Martina Cardillo and A. Carosi and Marie-Sophie Carrasco and F. Cassol-Brunner and D. Cauz and Davide Cerasole and Giovanni Ceribella and Yating Chai and Ks Cheng and A. Chiavassa and Michiyuki Chikawa and L. Chytka and Alejo Cifuentes and J. L. Contreras and J. Cortina and H. Costantini and M. Dalchenko and F. Dazzi and Alessandro De Angelis and Mathieu de Bony de Lavergne and B. De Lotto and Mario de Lucia and Raniere de Menezes and L. del Peral and Guillaume Deleglise and C. Delgado Mendez and J. Delgado and Domenico Della Volpe and Michael Dellaiera and Ambra Di Piano and F. Di Pierro and A. Di Pilato and Riccardo Di Tria and L. Di Venere and C. Díaz and Rune Michael Dominik and D. Dominis Prester and Alice Donini and D. Dorner and M. Doro and Laura Eisenberger and Dominik Elsässer and Gabriel Emery and Juan Escudero and V. Fallah Ramazanı and Giovanna Ferrara and F. Ferrarotto and A. Fiaßon and L. Foffano and Lluís Freixas Coromina and Stefan Fröse and Satoshi Fukami and Yasushi Fukazawa and Enrique Rivera Garcí­a and R. J. Garcı́a López and Claudio Gasbarra and Dario Gasparrini and Felix Geyer and Joao Giesbrecht Paiva and N. Giglietto and F. Giordano and Paweł Gliwny and N. Godinović and Roger Grau Haro},
  title = {{The Real Time Analysis framework of the Cherenkov Telescope Array's Large-Sized Telescope}},
  year = {2023},
  doi = {10.22323/1.444.0616},
}

Loading related content...

Figure