記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

LEOPARD (Light intensity Experiment with On-orbit Positioning and satellite Ranging Demonstration) is a 3-unit (3U) research CubeSat with various mission objectives such as light-scattering observation over the horizon with a multispectral camera, onboard processing of Earth-origin one-way radio ranging signal (OPERA), single event latch-up (SEL) detection, solar panel deployment demonstration with shape memory alloy (SMA), and measurement of magnetic field independent components of stray and natural fields. The solar panel deployment mechanism utilizes shape memory alloy with a heater to ensure controlled panel deployment. In this paper, the structural design of the LEOPARD CubeSat is presented in addition to the assembly and integration procedures. Furthermore, structural analysis is presented from modal analysis to static load analysis, and the simulation of stress distributions and validation of structural integrity have been performed with finite element analysis. Vibration testing is also conducted to evaluate the system level response to mechanical excitations during launch, ensuring robustness against dynamic loads. LEOPARD used a slot-type design for subsystem integration that has been used in our previous satellites, and the novelty of the structure design comes from the deployment mechanism and method. Finally, the LEOPARD flight model is under testing, and the operation is expected to begin in the second half of 2025.

DOI： 10.52202/078379-0039

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85218458171&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Due to advances in observation and imaging technologies, modern astronomical satellites generate large volumes of data. This necessitates efficient onboard data processing and high-speed data downlink. Reflecting this trend is the Visible Extragalactic background RadiaTion Exploration by CubeSat (VERTECS) 6U Astronomical Nanosatellite. Designed for the observation of Extragalactic Background Light (EBL), this mission is expected to generate a substantial amount of image data, particularly within the confines of CubeSat capabilities. This paper introduces the VERTECS Camera Control Board (CCB), an open-source payload interface board leveraging Commercial Off-The-Shelf (COTS) components, with a Raspberry Pi Compute Module 4 at its core. The VERTECS CCB hardware and software have been designed from the ground up to serve as the sole interface between the VERTECS bus system and astronomical imaging payload, while providing compute capability not usually seen in nanosatellites of this class. Responsible for mission data processing, it will facilitate high-speed data transfer from the imaging payload via gigabit Ethernet, while also providing a high-bitrate serial connection to the payload X-band transmitter for mission data downlink. Additional interfaces for secondary payloads are provided via USB-C and standard 15-pin camera connectors. The Raspberry Pi embedded within the VERTECS CCB operates on a standard Linux distribution, streamlining the software development process. Beyond addressing the current mission’s payload control and data handling requirements, the CCB sets the stage for future missions with heightened data demands. Furthermore, it supports the adoption of machine learning and other compute-intensive applications in orbit. This paper delves into the development of the VERTECS CCB, offering insights into the design and validation of this next-generation payload interface, to ensure that it can survive the rigors of space flight.

DOI： 10.1117/12.3019471

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85201968106&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

The design of the attitude determination and control system (ADCS) of CubeSats plays a crucial role in the success of their missions. ADCS design is a challenging task for CubeSats with commercial-off-the-shelf, low-cost, and small-size equipment due to limited resources. In this paper, the proposed ADCS design of the LEOPARD satellite is presented. The LEOPARD is a 3U cubesat developed by the Kyushu Institute of Technology and its main missions are to perform technology demonstration for an on-orbit positioning system and observe the horizon for light-intensity experiment. The light-intensity experiment requires 8 degree attitude control accuracy to realize its mission. The ADCS of LEOPARD is equipped with two three-axis gyroscopes, two three-axis magnetometers, six coarse sun sensors, a three-axis magnetorquer, and a y-axis reaction wheel. The attitude determination part consists of a coarse attitude determination algorithm that is to be used to initialize the attitude information and an extended Kalman filter as a main attitude estimation technique that makes use of all the sensors. As a backup solution, a gyroless extended Kalman filter is designed in case of gyro failure. An estimation algorithm is designed to determine the residual magnetic moment. The attitude control algorithms provide the satellite detumbling using the B-dot control algorithm and target-pointing PD-type control algorithms. The estimated residual magnetic moment is also compensated by the feedforward control approach. The simulations show that the attitude and angular velocity could be estimated in the Sun phase as well as in eclipse using gyroscope measurements in addition to the magnetometer and sun sensor. Nadir pointing and Sun pointing modes can be realized using solely magnetic actuation within the requirements for communication and Sun acquisition. Horizon detection algorithm with Sun pointing is achieved for scientific missions in sunrise and sunset phases by using a reaction wheel in addition to magnetorquers.

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85187984917&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

The KITSUNE satellite is a 6-unit CubeSat platform with the main mission of 5-m-class Earth observation in low Earth orbit (LEO), and the payload is developed with a 31.4 MP commercial off-the-shelf sensor, customized optics, and a camera controller board. Even though the payload is designed for Earth observation and to capture man-made patterns on the ground as the main mission, a secondary mission is planned for the classification of wildfire images by the convolution neural network (CNN) approach. Therefore, KITSUNE will be the first CubeSat to employ CNN to classify wildfire images in LEO. In this study, a deep-learning approach is utilized onboard the satellite in order to reduce the downlink data by pre-processing instead of the traditional method of performing the image processing at the ground station. The pre-trained CNN models generated in Colab are saved in RPi CM3+, in which, an uplink command will execute the image classification algorithm and append the results on the captured image data. The on-ground testing indicated that it could achieve an overall accuracy of 98% and an F1 score of a 97% success rate in classifying the wildfire events running on the satellite system using the MiniVGGNet network. Meanwhile, the LeNet and ShallowNet models were also compared and implemented on the CubeSat with 95% and 92% F1 scores, respectively. Overall, this study demonstrated the capability of small satellites to perform CNN onboard in orbit. Finally, the KITSUNE satellite is deployed from ISS on March 2022.

DOI： 10.3390/rs14081874

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85129121668&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In the past decades, deep learning (DL) has been a powerful tool for image recognition. Any image acquired by a camera sensor or a satellite camera payload can now be rapidly classified using DL. At present, satellite imagery is being used to observe the weather and map natural disasters which occur globally, such as forest fires. On the other hand, high-resolution forest fire images from defined datasets are found to be difficult to obtain and be trained in DL models. In this study, a new forest fire dataset was collected from PlanetScope, Sentinel-2, and Landsat-8 and was pre-trained using convolution neural networks (CNN) models to create a GUI called CoFFI: a Classification of Forest Fire Imagery GUI in order to post-processing the image downlink from satellites. Four labels (cloud, land, sea, and wildfire) were trained using the existing networks and analyzed for accuracy and computational time. The training results showed the highest accuracy of 99%, with a 97% F1-score of wildfire labels utilizing the ResNet architecture. Other models were further evaluated and compared to show the effectiveness of the dataset created. Furthermore, these networks were also included in the CoFFI GUI as options to predict new images downloaded from satellites. The GUI was partially tested and verified for images acquired from KITSUNE 6U CubeSat, which captured non-forest fire images. Ultimately, this computer program would be valuable for other satellite projects by filtering first the enormous image data downlink to the ground station before providing it to the respective authority.

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85167578906&origin=inward

担当区分：筆頭著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Developing star trackers quickly is non-trivial. Achieving reproducible results and comparing different algorithms are also open problems. In this sense, this work proposes the use of synthetic star images (a simulated sky), allied with the standardized structure of the Universal Verification Methodology as the base of a design approach. The aim is to organize the project, speed up the development time by providing a standard verification methodology. Future rework is reduced through two methods: a verification platform that us shared under a free software licence; and the layout of Universal Verification Methodology enforces reusability of code through an object-oriented approach. We propose a black-box structure for the verification platform with standard interfaces, and provide examples showing how this approach can be applied to the development of a star tracker for small satellites, targeting a system-on-a-chip design. The same test benches were applied to both early conceptual software-only implementations, and later optimized software-hardware hybrid systems, in a hardware-in-the-loop configuration. This test bench reuse strategy was interesting also to show the regression test capability of the developed platform. Furthermore, the simulator was used to inject specific noise, in order to evaluate the system under some real-world conditions.

DOI： 10.3390/s21030907

Scopus

CiNii Article

CiNii Research

PubMed

その他リンク： https://kyutech.repo.nii.ac.jp/records/6903

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

In the past decade, a massive number of wildfire events have been reported worldwide. An economic loss every year shows the importance of having satellites to overcome this catastrophe. Earth observation CubeSat can be a solution to detect, monitor, and provide data for the fire departments from the aerial view. For this purpose, the utilization of the deep learning (DL) algorithm to process the images captured onboard CubeSat before they are downlinked to the ground station is demonstrated. As a proof-of-concept, a single-board computer (SBC), Raspberry Pi, has been integrated into the KITSUNE 6U CubeSat to run the model. In this paper, the results of functional and environment tests of our proof-of-concept implementation have shown the feasibility of using image classification onboard the CubeSat. The DL algorithm's accuracy has reached 95% in the ground test, and this accuracy can be improved with further verifications of the platform and analysis of flight results. Therefore, DL's advancement in CubeSat is the state-of-art that can contribute significantly to address the wildfire issues.

DOI： 10.1109/IGARSS47720.2021.9553108

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85126019681&origin=inward

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）

Wildfires burn millions of hectares of land every year globally. Most of them are caused by humans, while only 10-15% occur naturally due to the climate change. The hotter weather dries out forests and plants, making them more prone to fire. The "frontline wildfire defense" has fully utilized satellite imagery to monitor, map, and control the fire spread and damage. However, there are three major challenges of using traditional satellite data: (1) the spatial resolution, (2) the temporal resolution, and (3) the downlink and analyzing data on the ground. In recent technology, the satellites are developed into small-size CubeSats that supporting the resolution issues. By exploiting the deep learning (DL) technique, the CubeSat can become sufficiently "intelligent" to detect wildfire events. This paper discusses a potential approach for implementing a Convolution Neural Network (CNN) onboard a CubeSat to sense wildfire. The DL model has been tested on the Camera Controller Board (CCB) embedded with Raspberry Pi Compute Module (RPi CM3+), that interfacing with the imaging mission of a 6U CubeSat named KITSUNE. In addition, the space environment test of radiation Total Ionizing Dose (TID) with functional tests of the board has been discussed. The results have shown no anomaly observed on the RPi while the DL model achieved a 94% overall accuracy with 16 minutes of learning time and 32 seconds of classification time. Hence, the state-of-art processing images onboard CubeSat will improve the valuable downlink data as the limited time window passes through the ground station.

DOI： 10.1117/12.2603906

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85122980273&origin=inward

記述言語：英語   掲載種別：研究論文（学術雑誌）

The execution of centroid extraction algorithms using a microprocessor consumes considerable resources when compared to the other steps involved in star trackers. This paper presents a method to identify star centroids in star trackers by pre-processing the pixels using a field-programmable gate array (FPGA) directly in the stream transmitted by an image sensor. The dedicated hardware filters the star pixels and transmits them to a processor, which computes the centroids of the respective image using an infinite impulse response filter. Thus, there is a substantial decrease in memory consumption and a reduction of the processor usage during the attitude determination computation, making the process more attractive for small satellites. A hardware-in- the-loop simulation is presented to test the performance of the system. It was possible to achieve a subpixel precision in the centroid coordinates' estimation, and also lower execution times in comparison with methods based on the processing of whole images.

DOI： 10.1504/IJSNET.2020.104458

Kyutacar

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85078035744&origin=inward