The radiation environment in space can pose a serious risk to both humans and space systems. Widespread and continuous monitoring of this environment is essential to mitigate risks associated with radiation exposure. Miniaturization and use of commercial-off-the-shelf components have enabled significant advances in space technology. These trends can be leveraged to develop innovative radiation sensing and monitoring technologies. However, dosimeters that can effectively measure radiation levels while minimizing their impact on size, power, mass, and cost are required. Floating gate dosimeters (FGDOSs) possess these characteristics, but rigorous testing is needed to ensure their accuracy in spacecraft applications. In this study, we conducted an extensive characterization campaign for an FGDOS chip using a proton beam, increasing the available information on the sensor. The behavior of the dosimeter with respect to resolution, dose rate, beam energy, total ionizing dose (TID), power consumption, annealing, temperature, and single-event effects (SEEs) was experimentally studied. Notably, we observed a previously unseen phenomenon, which we termed 'frequency surge' (FS). This phenomenon is likely to have implications for the dosimeter's performance under real spacecraft conditions. Our findings show that the dosimeter is able to combine small power consumption with high dose resolution but also highlight the need for testing against other radiation source types and intensities.

@en

This chapter provides an overview of the command and data handling system (CDHS) in small satellites and CubeSats. The chapter presents first analysis of radiation effects, specifically targeted at this subsystem, to justify components and architecture choices. Improvements in radiation testing strategies are also presented, specifically for small satellites. State-of-the-art components are then presented, providing an overview of the current market and the most common architectures. An overview of past and current missions is also presented, providing a clear mapping of the presented state-of-the-art components and architectures to guide future designs. High-level design considerations are also presented to help the reader follow some of the current trends in the sector. This chapter, overall, aims at presenting the most common approaches for the CDHS system and comparing this with traditional satellites, showing where the main differences lay with component selection and testing strategies being the fundamental points driving the architecture choices.

@en

The last years saw the diffusion of nano, pico and femto satellite missions launched by multiple entities thanks to the launch cost reduction and the electronics miniaturization. Such missions usually present limited capabilities in terms of precise orbit determination and extremely small radar and optical cross-sections. Often these missions carry one or more laser retro-reflectors for precise orbit determination but precise orbital measurements cannot be found in the literature. Miniaturized GNSS receivers are also often carried out but due to the experimental nature of such missions, the reliability and time span of such measurements is limited, leaving radar tracking as the only reliable tracking method. Due to the size of such satellites, the signal-to-noise ratio of such radar measurements is typically low and satellite identification (when launched on ride-share launches with a hundred or more other satellites) proves difficult and time-consuming. Being these very small satellites at the edge of the radar detection capabilities and not providing independent orbit determination means, their position uncertainty could be quite significant, leading to an increased orbit collision perceived risk. With this paper, we present a dedicated small satellite formation, made by multiple nano and pico satellites to evaluate the space surveillance network tracking capabilities and limits. The formation is made by a 3U CubeSat to be deployed as part of a rideshare launch. The satellite would be equipped with multiple means to track it, including a GNSS receiver, a set of multiple laser retro-reflectors, and LEDs for optical, laser, and radar tracking, allowing to characterize also different detection means in terms of capabilities. Such a satellite is made of two independent smaller satellites that can be un-docked in orbit upon command, reducing the satellite size and cross-section. This would push the detection limit for the space surveillance networks starting from an already acquired object and with limited clutter around it. Independent laser and GNSS tracking would allow ground measurement validation and validate position estimations. Further pico-satellites would be deployed by each sub-satellite to further push the detection limits and validate up to which size objects are trackable (still optically, radar and GNSS), thanks to miniaturized GNSS receivers already flown by several other missions. Sub-satellite separation is implemented upon command to ensure the process can be followed and executed at lower altitudes to limit the orbital lifetime of eventually hard-to-track small objects that could worsen the space debris problem. Ground characterization (in terms of optical and radar properties) will be performed, also including polarimetric measurements used to identify the separate satellites. All these technologies together would contribute to creating a unique tool to estimate the tracking capabilities of multiple instruments, specifically tailored for very small objects, the hardest to track, as compared to other characterization activities performed on much bigger objects.@en

This paper describes the work carried out to extend the NOEL-V platform to include data-level parallelism (DLP) by implementing an integer subset of the RISC-V Vector Extension. The performance and resource utilization efficiency of the resulting vector processor for different levels of DLP (i.e., number of lanes) have been compared to the baseline scalar processor on a Xilinx Kintex Ultrascale field-programmable gate array, employing typical kernels for compute-intensive applications. The role of the memory subsystem has also been investigated, comparing the results obtained with a low-latency and a high-latency main memory. The results show that the speed-up due to the use of the vector pipeline increases with the number of lanes in the vector processor, achieving up to 23.0× the performance of the scalar processor with only 4.3× the resources of the baseline scalar processor. Using an implementation with 32 lanes increases performance even for problem sizes larger than the number of lanes, achieving up to more than 11.7× the performance of the scalar processor with just 1.9× its resource utilization for 128 × 128 matrix multiplications. This work proves that implementations of the selected subset are easily scalable and fit for small-processor implementations in highly constrained space embedded systems.

@en

This paper introduces an adaptive Convolutional Neural Network (CNN)-based Unscented Kalman Filter for the pose estimation of uncooperative spacecraft. The validation is carried out at Stanford's robotic Testbed for Rendezvous and Optical Navigation on the Satellite Hardware-In-the-loop Rendezvous Trajectories (SHIRT) dataset, which simulates vision-based rendezvous trajectories of a servicer spacecraft to PRISMA's Tango spacecraft. The proposed navigation system is stress-tested on synthetic as well as realistic lab imagery by simulating space-like illumination conditions on-ground. The validation is performed at different levels of the navigation system by first training and testing the adopted CNN on SPEED+, Stanford's spacecraft pose estimation dataset with specific emphasis on domain shift between a synthetic domain and an Hardware-In-the-Loop domain. A novel data augmentation scheme based on light randomization is proposed to improve the CNN robustness under adverse viewing conditions, reaching centimeter-level and 10 degree-level pose errors in 80% of the SPEED+ lab images. Next, the entire navigation system is tested on the SHIRT dataset. Results indicate that the inclusion of a new scheme to adaptively scale the heatmaps-based measurement error covariance based on filter innovations improves filter robustness by returning centimeter-level position errors and moderate attitude accuracies, suggesting that a proper representation of the measurements uncertainty combined with an adaptive measurement error covariance is key in improving the navigation robustness.

@en

The Delft University of Technology has been working on Delfi-PQ, a 3 P P ocketQube d eveloped b y Aerospace Engineering students during their education. The satellite, while being only 50x50x178 mm and having a mass of 545 g, shares the same problems and requirements of bigger satellites. This paper presents the design concept, development, and testing of Delfi-PQ to help other teams in their development. All the combined information will help to generate a big picture for institutions to start their own small satellite mission.@en

The Earth-Moon system is constantly bombarded by meteoroids of different size and impact speed. Observation of the impacts on the Moon can enable thorough characterization of the Lunar meteoroid flux, which is similar to that of the Earth. While Earth-based Lunar observations are restricted by weather, geometric and illumination conditions, a Lunar-based observation campaign can improve the detection rate and, when observing the Lunar far side, complement in both space and time the observations taken from Earth. The Lunar Meteoroid Impact Observer (LUMIO), one of the two winning concepts of the ESA SysNova Lunar CubeSats for Exploration challenge, is a mission designed to observe, quantify, and characterize the micro-meteoroid impacts on the Lunar far side. It is based on a 12U CubeSat that carries the LUMIO-Cam, a custom-designed optical instrument capable of detecting light flashes in the visible spectrum. The spacecraft is placed on a halo orbit about the Earth–Moon L2 point, where permanent full-disk observation of the Lunar far side can be performed with excellent quality, given the absence of Earth background noise. After passing Phase 0 and an independent feasibility study in the ESA Concurrent Design Facility, the mission has successfully completed its Phase A in March 2021. Although the Phase 0 design of the LUMIO spacecraft was assessed as feasible by the ESA CDF study, a number of critical issues were identified, which have been tackled by the Phase A design. The paper presents the outcome of this Phase A design effort for the LUMIO spacecraft. Particularly relevant changes or updates in the spacecraft design include: a consolidated design of the LUMIO-Cam, with longer baffle for straylight protection; a set of ADCS sensors and actuators with increased redundancy; a combination of Direct-to-Earth communication and inter-satellite link with a mothership in Lunar orbit; use of Earth ranging to complement and validate the current innovative autonomous navigation strategy based on optical observations of the Moon by means of the LUMIO-Cam; re-assessment of the COTS components selection for the power and propulsion systems.

@en

The estimation of the relative pose of an inactive spacecraft by an active servicer spacecraft is a critical task for close-proximity operations, such as In-Orbit Servicing and Active Debris Removal. Among all the challenges, the lack of available space images of the inactive satellite makes the on-ground validation of current monocular camera-based navigation systems a challenging task, mostly due to the fact that standard Image Processing (IP) algorithms, which are usually tested on synthetic images, tend to fail when implemented in orbit. In response to this need to guarantee a reliable validation of pose estimation systems, this paper presents the most recent advances of ESA's GNC Rendezvous, Approach and Landing Simulator (GRALS) testbed for close-proximity operations around uncooperative spacecraft. The proposed testbed is used to validate a Convolutional Neural Network (CNN)-based monocular pose estimation system on representative rendezvous scenarios with special focus on solving the domain shift problem which characterizes CNNs trained on synthetic datasets when tested on more realistic imagery. The validation of the proposed system is ensured by the introduction of a calibration framework, which returns an accurate reference relative pose between the target spacecraft and the camera for each lab-generated image, allowing a comparative assessment at a pose estimation level. The VICON Tracker System is used together with two KUKA robotic arms to respectively track and control the trajectory of the monocular camera around a scaled 1:25 mockup of the Envisat spacecraft. After an overview of the facility, this work describes a novel data augmentation technique focused on texture randomization, aimed at improving the CNN robustness against previously unseen target textures. Despite the feature detection challenges under extreme brightness and illumination conditions, the results on the high exposure scenario show that the proposed system is capable of bridging the domain shift from synthetic to lab-generated images, returning accurate pose estimates for more than 50% of the rendezvous trajectory images despite the large domain gaps in target textures and illumination conditions.

@en

The Lunar Meteoroid Impacts Observer (LUMIO) is a CubeSat mission to observe, quantify, and characterize the meteoroid impacts on the lunar farside by detecting their flashes. This complements the knowledge gathered by Earth-based observations of the lunar nearside, thus synthesizing global information on the lunar meteoroid environment and contributing to the lunar situational awareness. The goal of LUMIO is to advance our current knowledge of meteoroid models in the solar system. In this work, we present the methodology devised to predict the scientific contribution of LUMIO. Our approach relies on combined modeling and simulation of payload, orbit, and environment. The analyses carried out have been used to drive the design of the LUMIO mission and its payload, the LUMIO-Cam. A payload radiometric model is derived and exploited to assess the quality of the scientific measurements. A dedicated study about straylight rejection is carried out to assess how straylight noise affects LUMIO-Cam measurements. Our results indicate that a 150 mm baffle grants good performance when the Sun angle is between 20° and 90°. Furthermore, the present-day LUMIO mission has the potential to detect more than 6000 impact flashes during the activity peak of the Geminids in 2024 in the range of the equivalent impact kinetic energy at Earth of [10 ⁻⁶,10 ⁻¹]kton TNT Equivalent. Compared to previous programmes, LUMIO could refine information and fill the knowledge gap about the meteoroid population in the ranges of the equivalent impact kinetic energy at Earth of [10 ⁻⁶,10 ⁻⁴]kton TNT Equivalent and [10 ⁻⁴,10 ⁻¹]kton TNT Equivalent, respectively.

@en

The Lunar Meteoroid Impact Observer (LUMIO) is a CubeSat mission at the Earth-Moon Lagrangian point 2 (L2) designed to observe, quantify, and characterize the meteoroid impacts by detecting their flashes on the Lunar farside. LUMIO can be deployed as one of the payloads in the NASA Commercial Lunar Payload System or from Artemis-2 mission to a low Lunar orbit and to demonstrate autonomous navigation capabilities to reach its operational orbit around the Earth-Moon L2. From there, its scientific mission to map and investigate the spatial and temporal characteristics of meteoroids impacting the Lunar surface will start and is expected to last for one year. LUMIO is a 12U CubeSat including a dedicated camera to monitor impact flashes in the visible and near-infrared spectrum, and also allows estimating the impact of temperature and energy. Optical navigation using the payload camera will also demonstrate increased on-board autonomy and drastically reduced mission costs. Navigation validation will be carried out using standard ground-based radiometric techniques enabled by a miniaturized X-band coherent transponder on-board. LUMIO can also use an inter-satellite link for telemetry and control via a commercial Lunar data relay system, providing a redundant communication system and lowering the need for high-gain ground stations for routine operations. The satellite bus derives from a commercial version designed for Low Earth Orbit and it will feature several improvements to operate in the Lunar environment, including a more advanced thermal control and radiation shielding. Commercial Off-The-Shelf systems will require a radiation screening and this will contribute to maintain the mission budget low and aim at a launch date in 2024.@en

Soft errors in embedded systems' memories like single-event upsets and multiple-bit upsets lead to data and instruction corruption. Therefore, devices deployed in harsh environments, such as space, use fault-tolerant processors or redundancy methods to ensure critical application dependability. Another rising concern in secure, critical space applications is the possible introduction of hardware Trojans in an untrusted phase of the manufacturing process. Besides environmental side-effects, an adversary that has injected a malicious mechanism e.g., in the processor or memory can trigger unwanted behavior or leak sensitive information. Techniques to prevent or mitigate hardware Trojans are important to ensure hardware security. Leveraging the openness of the RISC-V ISA, this paper introduces a novel solution to improve the security and dependability of softcores with a low area and latency overhead. The instruction validator which is the first part of this solution can effectively detect hardware Trojans and multiple-bit upsets in the instruction memory by checking instruction/address pairs using a Bloom filter probabilistic data structure. The second part of the solution is the proposal of an error correction code instruction memory using Hamming single-error correction to detect and correct single-event upsets. It has also been proven that the Hamming decoder improves the detection performance of the instruction validator.

@en

RISC-V is an open and modular Instruction Set Architecture(ISA) which is rapidly growing in popularity in terrestrial applications. This paper presents the place in future space embedded systems ESA's roadmap for RISC-V based processors. In order to satisfy different applications with contrasting requirements in satellite data systems, four different types of processors are identified: 1) General-Purpose (GP) processors for payloads 2) main platform On-Board Computers (OBCs) controllers 3) low-area/low-power microcontrollers (uCs), 4) enhanced payload processors with support for Artificial Intelligence (AI). We also describe the state of the art of the RISC-V software ecosystem, including the currently available hardware platforms, with a focus on developments for space applications and what has already been done in the European Space Industry. Finally, planned activities are presented, with a focus on the role of the European ecosystem.@en

Integrated circuits employed in space applications generally have very low-volume production and high performance requirements. Therefore, the adoption of Commercial-Off-The-Shelf (COTS) components and Third Party Intellectual Property cores (3PIPs) is of extreme interest to make system design, implementation and deployment cost-effective and viable w.r.t. performance. On the other hand, this design paradigm exposes the system to a number of security threats both at design-time and at runtime. In this paper, we discuss the security issues related to space applications mainly focusing on threats that come from the adoption of the well-known RISCV microprocessor. We highlight how Hardware Trojan horses (HTHs) and Microarchitectural Side-Channel Attacks (MSCAs) may compromise the overall system operation by either altering its nominal behavior or by stealing secret information. We discuss the security extensions provided by the RISC-V architecture as well as their limitations. The paper is concluded by an overview of the issues that are still open regarding the security of such microprocessor in the space domain.

@en

The objective of this paper is to investigate which approach would lead to more reliable CubeSats: full subsystem redundancy or improved testing. Based on data from surveys, the reliability of satellites and subsystems is estimated using a Kaplan–Meier estimator. Subsequently, a variety of reliability models is defined and their maximum likelihood estimates are compared. A product of a Lognormal distribution addressing immaturity failure and a Gompertz distribution addressing wear-out is found to best represent CubeSat reliability. Bayesian inference is used to find realistic wear-out parameters by using failure data of small satellites. Subsystem reliability estimates are subsequently found using a similar approach. A reliability model for CubeSats with redundant subsystems is established, verified and applied in a Monte Carlo simulation. The results are compared with a model for reduced immaturity failure. Allocating resources to reduction of immaturity failures through improved testing is considered to be superior to allocating these resources to the implementation of subsystem redundancy.@en

The Lunar Meteoroid Impact Observer (LUMIO) is a mission designed to observe, quantify, and characterize the impacts of meteoroids on the lunar far side and, therefore, complement in both space and time the observations currently taken from Earth. While Earth-based lunar observations are restricted by weather, geometric and illumination conditions, a Moon-based observation campaign can improve the detection rate of impact flashes and the general quality and reliability of the final scientific product. The mission has successfully completed the Phase A design, after successfully passing Phase 0 and an independent study in the ESA Concurrent Design Facility that has fully confirmed its feasibility. The LUMIO spacecraft is a 12U XL CubeSat with a mass of around 28 kg, released into Lunar orbit by a carrier spacecraft and capable of autonomously transferring from this initial parking orbit to its final destination, a halo orbit about the Earth–Moon L2 point from which permanent fulldisk observation of the Lunar far side can be performed. The mission objectives will be achieved thanks to the LUMIO-Cam, a custom-designed optical instrument capable of detecting light flashes in the visible spectrum, and an innovative on-board data processing system, capable of drastically reducing the amount of information that needs to be transmitted back to Earth. The camera is capable of generating 5TBd−1 of data, out of which only approximately 1MBd−1 will need to be transmitted to Earth, since impact identification will be performed autonomously onboard and only relevant information will be actually transmitted. This paper will present the current status of the mission, summarising the main results of the Phase A design and the way forward to the following steps in mission implementation (Phases B-C). The paper will include a summary of the spacecraft system design and mission analysis. Particular focus will be given to the mission operations aspects. @en

The Lunar Meteoroid Impact Observer (LUMIO) is a mission designed to observe, quantify, and characterize the meteoroid impacts by detecting their flashes on the lunar far side. Earth-based lunar observations are restricted by weather, geometric and illumination conditions, while a lunar orbiter can improve the detection rate of lunar meteoroid impact flashes, as it would allow for longer monitoring periods. This paper will focus on the communications and radio navigation system of the mission, designed for the ESA roadmap for lunar exploration. LUMIO has been designed to operate autonomously after deployment from a lunar mother spacecraft in a low inclination lunar orbit and to reach without human intervention his final destination orbit close to the Earth-Moon L2 point, where science can be carried out. Being the destination orbit always in view from Earth (despite a distance of 460000 - 480000 km), Direct-to-Earth communication was added to the mission as a mean to reduce risk and allow independent verification of several of the innovative technologies that would be demonstrated, first of all autonomous navigation. A detailed link budget analysis will be presented for all mission phases for both the link with the mother spacecraft in low lunar orbit and the link with Earth. Beside defining the achievable data transfer, we will focus also on evaluating the available ground stations to better evaluate mission cost with respect to science return. Radio-navigation performances will also be evaluated to estimate the position and relative velocity accuracy, given also the limited performances available for the on-board navigation transponder. This will help also better defining the on-board autonomous navigation system, constraining the total error budget. Further strategies, such as beacon tones, will be evaluated to lower the overall operational cost by employing continuous monitoring with a low performances ground station and, only when needed, perform high speed downlink using a deep-space class ground station. This strategy is considered of extreme importance, especially for small missions, to allow opportunistic operations on high gain antennas, given their very busy schedule. Keywords: LUMIO, CubeSat, Lunar, Radio, link @en

The Earth-Moon system is constantly bombarded by meteoroids of different size and impact speed. Observation of the impacts on the Moon can enable thorough characterization of the Lunar meteoroid flux, which is similar to that of the Earth. While Earth-based Lunar observations are restricted by weather, geometric and illumination conditions, a Lunar-based observation campaign can improve the detection rate and, when observing the Lunar far side, complement in both space and time the observations taken from Earth. The Lunar Meteoroid Impact Observer (LUMIO), one of the two winning concepts of the ESA SysNova Lunar CubeSats for Exploration challenge, is a mission designed to observe, quantify, and characterize the micro-meteoroid impacts on the Lunar far side. It is based on a 12U CubeSat that carries the LUMIO-Cam, a custom-designed optical instrument capable of detecting light flashes in the visible spectrum. The spacecraft is placed on a halo orbit about the Earth–Moon L2 point, where permanent full-disk observation of the Lunar far side can be performed with excellent quality, given the absence of Earth background noise. After passing Phase 0 and an independent feasibility study in the ESA Concurrent Design Facility, the mission has successfully completed its Phase A in March 2021. Although the Phase 0 design of the LUMIO spacecraft was assessed as feasible by the ESA CDF study, a number of critical issues were identified, which have been tackled by the Phase A design. The paper presents the outcome of this Phase A design effort for the LUMIO spacecraft. Particularly relevant changes or updates in the spacecraft design include: a consolidated design of the LUMIO-Cam, with longer baffle for straylight protection; a set of ADCS sensors and actuators with increased redundancy; a combination of Direct-to-Earth communication and inter-satellite link with a mothership in Lunar orbit; use of Earth ranging to complement and validate the current innovative autonomous navigation strategy based on optical observations of the Moon by means of the LUMIO-Cam; re-assessment of the COTS components selection for the power and propulsion systems. @en

The Earth-Moon system is constantly bombarded by meteoroids of different size and impact speed. Observation of the impacts on the Moon can enable thorough characterization of the Lunar meteoroid flux, which is similar to that of the Earth. While Earth-based Lunar observations are restricted by weather, geometric and illumination conditions, a Lunar-based observation campaign can improve the detection rate and, when observing the Lunar far side, complement in both space and time the observations taken from Earth. The Lunar Meteoroid Impact Observer (LUMIO), one of the two winning concepts of the ESA SysNova Lunar CubeSats for Exploration challenge, is a mission designed to observe, quantify, and characterize the micro-meteoroid impacts on the Lunar far side. It is based on a 12U CubeSat that carries the LUMIO-Cam, a custom-designed optical instrument capable of detecting light flashes in the visible spectrum. The spacecraft is placed on a halo orbit about the Earth–Moon L2 point, where permanent full-disk observation of the Lunar far side can be performed with excellent quality, given the absence of Earth background noise. After passing Phase 0 and an independent feasibility study in the ESA Concurrent Design Facility, the mission has successfully completed its Phase A in March 2021. Although the Phase 0 design of the LUMIO spacecraft was assessed as feasible by the ESA CDF study, a number of critical issues were identified, which have been tackled by the Phase A design. The paper presents the outcome of this Phase A design effort for the LUMIO spacecraft. Particularly relevant changes or updates in the spacecraft design include: a consolidated design of the LUMIO-Cam, with longer baffle for straylight protection; a set of ADCS sensors and actuators with increased redundancy; a combination of Direct-to-Earth communication and inter-satellite link with a mothership in Lunar orbit; use of Earth ranging to complement and validate the current innovative autonomous navigation strategy based on optical observations of the Moon by means of the LUMIO-Cam; re-assessment of the COTS components selection for the power and propulsion systems. @en

The Lunar Meteoroid Impact Observer (LUMIO), one of the two winning concepts of the SysNova Lunar CubeSats for Exploration call by ESA, is a mission designed to observe, quantify, and characterize the meteoroid impacts on the Lunar far side by detecting the flashes generated by the impact. While Earth-based Lunar observations are restricted by weather, geometric and illumination conditions, a Lunar-based observation campaign can improve the detection rate and, when observing the Lunar far side, complement in both space and time the observations taken from Earth. The mission, which has successfully completed its Phase A in March 2021, is based on a 12U CubeSat that carries the LUMIO-Cam, a custom-designed optical instrument capable of detecting light flashes in the visible spectrum. The spacecraft is placed on a halo orbit about the Earth–Moon L2 point, where permanent full-disk observation of the Lunar far side can be performed with excellent quality, given the absence of background noise due to the Earth. The propulsion system is one of the most crucial design choices for the LUMIO spacecraft. It accomplishes various functions: orbital transfer from the initial Lunar orbit to the final halo orbit around L2, station keeping, reaction wheel desaturation, end of life disposal manoeuvres. The total required Delta-V budget for orbital transfer and station keeping is 201.8 m/s, plus an additional total impulse for reaction control tasks ranging from 110 Ns to 170 Ns, depending on the type of reaction control system that is selected. This paper presents a detailed summary of the phase A selection and design of the LUMIO propulsion system, based on the full list of requirements generated by the mission analysis. The main challenges of this process and the way they have been tackled are presented and discussed, including: use of two separate systems as opposed to an integrate one for main propulsion and reaction control tasks; availability of sufficiently reliable European propulsion options, to reduce the general mission costs; feasibility of replacing a chemical/cold gas system with electric propulsion; possible need for custom changes to the design of the selected COTS option (e.g. due to tank sizing). @en

The use of deep neural networks (DNNs) in terrestrial applications went from niche to widespread in a few years, thanks to relatively inexpensive hardware for both training and inference, and large datasets available. The applicability of this paradigm to space systems, where both large datasets and inexpensive hardware are not readily available, is more difficult and thus still rare. This paper analyzes the impact of DNNs on the system-level capabilities of space systems in terms of on-board decision making (OBDM) and identifies the specific criticalities of deploying DNNs on satellites. The workload of DNNs for on-board image and telemetry analysis is analyzed, and the results are used to drive the preliminary design of a RISC-V vector processor to be employed as a generic platform to enable energy-efficient OBDM for both payload and platform applications. The design of the memory subsystem is carried out in detail to allow full exploitation of the computational resources in typically resource-constrained space systems.

@en