-
Embedded Firmware Development for a Novel CubeSat Gamma-Ray Detector
Authors:
Joseph Mangan,
David Murphy,
Rachel Dunwoody,
Maeve Doyle,
Alexey Ulyanov,
Lorraine Hanlon,
Brian Shortt,
Sheila McBreen,
Masoud Emam,
Jessica Erkal,
Joe Flanagan,
Gianluca Fontanesi,
Andrew Gloster,
Conor O'Toole,
Favour Okosun,
Rakhi RajagopalanNair,
Jack Reilly,
Lána Salmon,
Daire Sherwin,
Joseph Thompson,
Sarah Walsh,
Daithí de Faoite,
Mike Hibbett,
Umair Javaid,
Fergal Marshall
, et al. (4 additional authors not shown)
Abstract:
The Gamma-ray Module (GMOD) is an experiment designed for the detection of gamma-ray bursts in low Earth orbit as the principal scientific payload on a 2-U CubeSat, EIRSAT-1. GMOD comprises a cerium bromide scintillator coupled to silicon photomultipliers which are processed and digitised by a bespoke ASIC. Custom firmware on the GMOD motherboard has been designed, implemented and tested for the M…
▽ More
The Gamma-ray Module (GMOD) is an experiment designed for the detection of gamma-ray bursts in low Earth orbit as the principal scientific payload on a 2-U CubeSat, EIRSAT-1. GMOD comprises a cerium bromide scintillator coupled to silicon photomultipliers which are processed and digitised by a bespoke ASIC. Custom firmware on the GMOD motherboard has been designed, implemented and tested for the MSP430 microprocessor which manages the experiment including readout, storage and configuration of the system. The firmware has been verified in a series of experiments testing the response over a realistic range of input detector trigger frequencies from 50Hz to 1kHz for the primary time tagged event (TTE) data. The power consumption and ability of the firmware to successfully receive and transmit the packets to the on-board computer was investigated. The experiment demonstrated less than 1% loss of packets up to 1kHz for the standard transfer mode with the power not exceeding 31mW. The transfer performance and power consumption demonstrated are within the required range of this CubeSat instrument.
△ Less
Submitted 4 November, 2021;
originally announced November 2021.
-
DAEMA: Denoising Autoencoder with Mask Attention
Authors:
Simon Tihon,
Muhammad Usama Javaid,
Damien Fourure,
Nicolas Posocco,
Thomas Peel
Abstract:
Missing data is a recurrent and challenging problem, especially when using machine learning algorithms for real-world applications. For this reason, missing data imputation has become an active research area, in which recent deep learning approaches have achieved state-of-the-art results. We propose DAEMA (Denoising Autoencoder with Mask Attention), an algorithm based on a denoising autoencoder ar…
▽ More
Missing data is a recurrent and challenging problem, especially when using machine learning algorithms for real-world applications. For this reason, missing data imputation has become an active research area, in which recent deep learning approaches have achieved state-of-the-art results. We propose DAEMA (Denoising Autoencoder with Mask Attention), an algorithm based on a denoising autoencoder architecture with an attention mechanism. While most imputation algorithms use incomplete inputs as they would use complete data - up to basic preprocessing (e.g. mean imputation) - DAEMA leverages a mask-based attention mechanism to focus on the observed values of its inputs. We evaluate DAEMA both in terms of reconstruction capabilities and downstream prediction and show that it achieves superior performance to state-of-the-art algorithms on several publicly available real-world datasets under various missingness settings.
△ Less
Submitted 30 June, 2021;
originally announced June 2021.
-
Anomaly Detection: How to Artificially Increase your F1-Score with a Biased Evaluation Protocol
Authors:
Damien Fourure,
Muhammad Usama Javaid,
Nicolas Posocco,
Simon Tihon
Abstract:
Anomaly detection is a widely explored domain in machine learning. Many models are proposed in the literature, and compared through different metrics measured on various datasets. The most popular metrics used to compare performances are F1-score, AUC and AVPR. In this paper, we show that F1-score and AVPR are highly sensitive to the contamination rate. One consequence is that it is possible to ar…
▽ More
Anomaly detection is a widely explored domain in machine learning. Many models are proposed in the literature, and compared through different metrics measured on various datasets. The most popular metrics used to compare performances are F1-score, AUC and AVPR. In this paper, we show that F1-score and AVPR are highly sensitive to the contamination rate. One consequence is that it is possible to artificially increase their values by modifying the train-test split procedure. This leads to misleading comparisons between algorithms in the literature, especially when the evaluation protocol is not well detailed. Moreover, we show that the F1-score and the AVPR cannot be used to compare performances on different datasets as they do not reflect the intrinsic difficulty of modeling such data. Based on these observations, we claim that F1-score and AVPR should not be used as metrics for anomaly detection. We recommend a generic evaluation procedure for unsupervised anomaly detection, including the use of other metrics such as the AUC, which are more robust to arbitrary choices in the evaluation protocol.
△ Less
Submitted 30 June, 2021;
originally announced June 2021.
-
Assembly, Integration, and Verification Activities for a 2U CubeSat, EIRSAT-1
Authors:
Sarah Walsh,
David Murphy,
Maeve Doyle,
Joseph Thompson,
Rachel Dunwoody,
Masoud Emam,
Jessica Erkal,
Joe Flanagan,
Gianluca Fontanesi,
Andrew Gloster,
Joe Mangan,
Conor O'Toole,
Favour Okosun,
Rakhi Rajagopalan Nair,
Jack Reilly,
Lána Salmon,
Daire Sherwin,
Paul Cahill,
Daithí de Faoite,
Umair Javaid,
Lorraine Hanlon,
David McKeown,
William O'Connor,
Kenneth Stanton,
Alexei Ulyanov
, et al. (2 additional authors not shown)
Abstract:
The Educational Irish Research Satellite, EIRSAT-1, is a project developed by students at University College Dublin that aims to design, build, and launch Ireland's first satellite. EIRSAT-1 is a 2U CubeSat incorporating three novel payloads; GMOD, a gamma-ray detector, EMOD, a thermal coating management experiment, and WBC, a novel attitude control algorithm. The EIRSAT-1 project is carried out w…
▽ More
The Educational Irish Research Satellite, EIRSAT-1, is a project developed by students at University College Dublin that aims to design, build, and launch Ireland's first satellite. EIRSAT-1 is a 2U CubeSat incorporating three novel payloads; GMOD, a gamma-ray detector, EMOD, a thermal coating management experiment, and WBC, a novel attitude control algorithm. The EIRSAT-1 project is carried out with the support of the Education Office of the European Space Agency, under the educational Fly your Satellite! programme. The Assembly, Integration and Verification plan for EIRSAT-1 is central to the philosophy and the development of the spacecraft. The model philosophy employed for the project is known as the 'prototype' approach in which two models of the spacecraft are assembled; an Engineering Qualification Model (EQM) and a Flight Model (FM). The payloads, GMOD and EMOD, and the Antenna Deployment Module (ADM) platform element warrant a Development Model in addition to an EQM and a FM, as they have been designed and developed in-house. After successful completion of the Critical Design Review and Ambient Test Readiness Review phases of the project, the EQM of EIRSAT-1 will be assembled and integrated. After assembly and integration of the EQM, the project will begin the ambient test campaign, in which the EQM undergoes ambient functional and mission testing. This work details the preparation and execution of the assembly, integration, and verification activities of EIRSAT-1 EQM.
△ Less
Submitted 20 October, 2020;
originally announced October 2020.
-
Flight Software Development for the EIRSAT-1 Mission
Authors:
Maeve Doyle,
Andrew Gloster,
Conor O'Toole,
Joseph Mangan,
David Murphy,
Rachel Dunwoody,
Masoud Emam,
Jessica Erkal,
Joe Flanaghan,
Gianluca Fontanesi,
Favour Okosun,
Rakhi Rajagopalan Nair,
Jack Reilly,
Lána Salmon,
Daire Sherwin,
Joseph Thompson,
Sarah Walsh,
Daithí de Faoite,
Umair Javaid,
Sheila McBreen,
David McKeown,
Derek O'Callaghan,
William O'Connor,
Kenneth Stanton,
Alexei Ulyanov
, et al. (2 additional authors not shown)
Abstract:
The Educational Irish Research Satellite, known as EIRSAT-1, is a student-led project to design, build, test and launch Ireland's first satellite. The on-board software for this mission is being developed using Bright Ascension's GenerationOne Flight Software Development Kit. This paper provides an overview of this kit and of EIRSAT-1's on-board software design. Drawing on the team's contrasting e…
▽ More
The Educational Irish Research Satellite, known as EIRSAT-1, is a student-led project to design, build, test and launch Ireland's first satellite. The on-board software for this mission is being developed using Bright Ascension's GenerationOne Flight Software Development Kit. This paper provides an overview of this kit and of EIRSAT-1's on-board software design. Drawing on the team's contrasting experience with writing entirely custom firmware for the mission's science payloads, this work discusses the impact of using a kit on the software development process. The challenges associated with the educational nature of this project are the focus of this discussion. The objective of this paper is to provide useful information for other CubeSat teams assessing software development options.
△ Less
Submitted 20 August, 2020;
originally announced August 2020.
-
Capturing Variabilities from Computed Tomography Images with Generative Adversarial Networks
Authors:
Umair Javaid,
John A. Lee
Abstract:
With the advent of Deep Learning (DL) techniques, especially Generative Adversarial Networks (GANs), data augmentation and generation are quickly evolving domains that have raised much interest recently. However, the DL techniques are data demanding and since, medical data is not easily accessible, they suffer from data insufficiency. To deal with this limitation, different data augmentation techn…
▽ More
With the advent of Deep Learning (DL) techniques, especially Generative Adversarial Networks (GANs), data augmentation and generation are quickly evolving domains that have raised much interest recently. However, the DL techniques are data demanding and since, medical data is not easily accessible, they suffer from data insufficiency. To deal with this limitation, different data augmentation techniques are used. Here, we propose a novel unsupervised data-driven approach for data augmentation that can generate 2D Computed Tomography (CT) images using a simple GAN. The generated CT images have good global and local features of a real CT image and can be used to augment the training datasets for effective learning. In this proof-of-concept study, we show that our proposed solution using GANs is able to capture some of the global and local CT variabilities. Our network is able to generate visually realistic CT images and we aim to further enhance its output by scaling it to a higher resolution and potentially from 2D to 3D.
△ Less
Submitted 29 May, 2018;
originally announced May 2018.
