Probing Convolutional Neural Networks for Event Reconstruction in γ-Ray Astronomy with Cherenkov Telescopes [IMA]

A dramatic progress in the field of computer vision has been made in recent years by applying deep learning techniques. State-of-the-art performance in image recognition is thereby reached with Convolutional Neural Networks (CNNs). CNNs are a powerful class of artificial neural networks, characterized by requiring fewer connections and free parameters than traditional neural networks and exploiting spatial symmetries in the input data. Moreover, CNNs have the ability to automatically extract general characteristic features from data sets and create abstract data representations which can perform very robust predictions. This suggests that experiments using Cherenkov telescopes could harness these powerful machine learning algorithms to improve the analysis of particle-induced air-showers, where the properties of primary shower particles are reconstructed from shower images recorded by the telescopes. In this work, we present initial results of a CNN-based analysis for background rejection and shower reconstruction, utilizing simulation data from the H.E.S.S. experiment. We concentrate on supervised training methods and outline the influence of image sampling on the performance of the CNN-model predictions.

Read this paper on arXiv…

T. Holch, I. Shilon, M. Buchele, et. al.
Mon, 20 Nov 17

Comments: 8 pages, 4 figures, Proceedings of the 35th International Cosmic Ray Conference (ICRC 2017), Busan, Korea

Development of a 32-channel ASIC for an X-ray APD Detector onboard the ISS [IMA]

We report on the design and performance of a mixed-signal application specific integrated circuit (ASIC) dedicated to avalanche photodiodes (APDs) in order to detect hard X-ray emissions in a wide energy band onboard the International Space Station. To realize wide-band detection from 20 keV to 1 MeV, we use Ce:GAGG scintillators, each coupled to an APD, with low-noise front-end electronics capable of achieving a minimum energy detection threshold of 20 keV. The developed ASIC has the ability to read out 32-channel APD signals using 0.35 $\mu$m CMOS technology, and an analog amplifier at the input stage is designed to suppress the capacitive noise primarily arising from the large detector capacitance of the APDs. The ASIC achieves a performance of 2099 e$^{-}$ + 1.5 e$^{-}$/pF at root mean square (RMS) with a wide 300 fC dynamic range. Coupling a reverse-type APD with a Ce:GAGG scintillator, we obtain an energy resolution of 6.7% (FWHM) at 662 keV and a minimum detectable energy of 20 keV at room temperature (20 $^{\circ}$C). Furthermore, we examine the radiation tolerance for space applications by using a 90 MeV proton beam, confirming that the ASIC is free of single-event effects and can operate properly without serious degradation in analog and digital processing.

Read this paper on arXiv…

M. Arimoto, S. Harita, S. Sugita, et. al.
Mon, 20 Nov 17

Comments: Accepted for publication in NIMA

Design of pre-optics for laser guide star wavefront sensor for the ELT [IMA]

In the present paper, we consider the optical design of a zoom system for the active refocusing in laser guide star wavefront sensors. The system is designed according to the specifications coming from the Extremely Large Telescope (ELT)-HARMONI instrument, the first-light, integral field spectrograph for the European (E)-ELT. The system must provide a refocusing of the laser guide as a function of telescope pointing and large decentring of the incoming beam. The system considers four moving lens groups, each of them being a doublet with one aspherical surface. The advantages and shortcomings of such a solution in terms of the component displacements and complexity of the surfaces are described in detail. It is shown that the system can provide the median value of the residual wavefront error of 13.8-94.3 nm and the maximum value <206 nm, while the exit pupil distortion is 0.26-0.36% for each of the telescope pointing directions.

Read this paper on arXiv…

E. Muslimov, K. Dohlen, B. Neichel, et. al.
Mon, 20 Nov 17

Comments: 7 pages, 5 figures, 3 tables

Exploring the brighter fatter effect with the Hyper Suprime-Cam [IMA]

The brighter fatter effect has been postulated to arise due to the build up of a transverse electric field, produced as photo-charges accumulate in the pixels’ potential wells. We investigate the brighter fatter effect in Hyper Suprime-Cam by examining flat fields and moments of stars. We observe deviations from the expected linear relation in the photon transfer curve, luminosity dependent correlations between pixels in flat field images and a luminosity dependent point spread function (PSF) in stellar observations. Under the key assumptions of translation invariance and Maxwell’s equations in the quasi-static limit, we give a first-principles proof that the effect can be parametrized by a translationally invariant scalar kernel. We describe how this kernel can be estimated from flat fields and discuss how this kernel has been used to remove the brighter fatter distortions in Hyper Suprime-Cam images. We find that our correction restores the expected linear relation in the photon transfer curves and significantly reduces, but does not completely remove, the luminosity dependence of the PSF over a wide range of magnitudes.

Read this paper on arXiv…

W. Coulton, R. Armstrong, K. Smith, et. al.
Mon, 20 Nov 17

Comments: 15 pages, 20 figures, submitted to the Astronomical Journal

Modeling UV Radiation Feedback from Massive Stars: I. Implementation of Adaptive Ray Tracing Method and Tests [IMA]

We present an implementation of an adaptive ray tracing (ART) module in the Athena hydrodynamics code that accurately and efficiently handles the radiative transfer involving multiple point sources on a three-dimensional Cartesian grid. We adopt a recently proposed parallel algorithm that uses non-blocking, asynchronous MPI communications to accelerate transport of rays across the computational domain. We validate our implementation through several standard test problems including the propagation of radiation in vacuum and the expansions of various types of HII regions. Additionally, scaling tests show that the cost of a full ray trace per source remains comparable to that of the hydrodynamics update on up to $\sim 10^3$ processors. To demonstrate application of our ART implementation, we perform a simulation of star cluster formation in a marginally bound, turbulent cloud, finding that its star formation efficiency is $12\%$ when both radiation pressure forces and photoionization by UV radiation are treated. We directly compare the radiation forces computed from the ART scheme with that from the M1 closure relation. Although the ART and M1 schemes yield similar results on large scales, the latter is unable to resolve the radiation field accurately near individual point sources.

Read this paper on arXiv…

J. Kim, W. Kim, E. Ostriker, et. al.
Mon, 20 Nov 17

Comments: 20 pages, 14 figures; accepted for publication in ApJ

Photogravimagnetic assists of light sails: a mixed blessing for Breakthrough Starshot? [IMA]

Upon entering a star system, light sails are subject to both gravitational forces and radiation pressure, and can use both in concert to modify their trajectory. Moreover, stars possess significant magnetic fields, and if the sail is in any way charged, it will feel the Lorentz force also.
We investigate the dynamics of so-called “photogravimagnetic assists” of sailcraft around $\alpha$ Centauri A, a potential first destination en route to Proxima Centauri (the goal of the Breakthrough Starshot program). We find that a 10m$^2$ sail with a charge-to-mass-ratio of around 10 $\mu$C/g or higher will need to take account of magnetic field effects during orbital maneouvres. The magnetic field can provide an extra source of deceleration and deflection, and allow capture onto closer orbits around a target star.
However, flipping the sign of the sailcraft’s charge can radically change resulting trajectories, resulting in complex loop-de-loops around magnetic field lines and essentially random ejection from the star system. Even on well-behaved trajectories, the field can generate off-axis deflections at $\alpha$ Centauri that, while minor, can result in very poor targetting of the final destination (Proxima) post-assist.
Fortunately for Breakthrough Starshot, nanosails are less prone to charging en route than their heavier counterparts, but can still accrue relatively high charge at both the origin and destination, when travelling at low speeds. Photogravimagnetic assists are highly non-trivial, and require careful course correction to mitigate against unwanted changes in trajectory.

Read this paper on arXiv…

D. Forgan, R. Heller and M. Hippke
Fri, 17 Nov 17

Comments: 10 pages, 8 figures, accepted for publication in MNRAS

Interstellar communication. III. Optimal frequency to maximize data rate [IMA]

The optimal frequency for interstellar communication, using “Earth 2017” technology, was derived in papers I and II of this series (arXiv:1706.03795, arXiv:1706.05570). The framework included models for the loss of photons from diffraction (free space), interstellar extinction, and atmospheric transmission. A major limit of current technology is the focusing of wavelengths $\lambda<300\,$nm (UV). When this technological constraint is dropped, a physical bound is found at $\lambda\approx1\,$nm ($E\approx\,$keV) for distances out to kpc. While shorter wavelengths may produce tighter beams and thus higher data rates, the physical limit comes from surface roughness of focusing devices at the atomic level. This limit can be surpassed by beam-forming with electromagnetic fields, e.g. using a free electron laser, but such methods are not energetically competitive. Current lasers are not yet cost efficient at nm wavelengths, with a gap of two orders of magnitude, but future technological progress may converge on the physical optimum. We recommend expanding SETI efforts towards targeted (at us) monochromatic (or narrow band) X-ray emission at 0.5-2 keV energies.

Read this paper on arXiv…

M. Hippke and D. Forgan
Fri, 17 Nov 17

Comments: N/A