Dr Chris Bridges

In this work, we propose a design space exploration workflow and tool for generating reconfigurable deep learning hardware models for FPGAs. The workflow is broken down into two main parts, Offline Design Exploration (ODE) and Online Design Reconfiguration ( ODR). Offline Design Exploration is automated through a workflow methodology which makes it possible for a designer to provide a Convolutional Neural Network (CNN) architecture and the option of providing additional design constraints in terms of latency and space. These automatically generate multiple design spaces which trade-off latency for resource utilization by dedicating varying processing elements through loop reordering, unrolling and pipelining. The second part of the process introduces online reconfiguration to the design space. Online Reconfiguration means the ability to modify the design at runtime after upload, by selectively running it partially or fully according to an application's immediate requirements, this provides the design with flexibility which, we believe, is highly beneficial for future autonomous on-board applications. We validate our work on the Xilinx Zynq-7100 board at 200 MHz and use custom trained networks architectures on three datasets for image classification, MNIST, SVHN and CIFAR-10. ODE generated designs achieved latency trade-offs of 95x for MNIST, 71x for CIFAR-10 and 18x for SVHN. Trade-offs in resource utilization in terms of DSP Slices were 44x for MNIST, 52x for SVHN and 24x for CIFAR-10. For the ODR, a 0.7% accuracy loss was traded-off with x13 speedup and a 25% reduction in power for MNIST, a 2% accuracy loss was traded-off with a 14x speedup and a 28% power reduction for SVHN, a 4% accuracy loss was traded off for a 50x speedup with 32.5% power reduction for CIFAR-10.

As megaconstellations are launched and the space sector grows, space debris pollution is posing an increasing threat to operational spacecraft. Low Earth orbit is a junkyard of dead satellites, rocket bodies, shrapnels, and other debris that travel at very high speed in an uncontrolled manner. Collisions at orbital speeds can generate fragments and potentially trigger a cascade of more collisions endangering the whole population, a scenario known since the late 1970s as the Kessler syndrome. In this work we present Kessler: an open-source Python package for machine learning (ML) applied to collision avoidance. Kessler provides functionalities to import and export conjunction data messages (CDMs) in their standard format and predict the evolution of conjunction events based on explainable ML models. In Kessler we provide Bayesian recurrent neural networks that can be trained with existing collections of CDM data and then deployed in order to predict the contents of future CDMs in a given conjunction event, conditioned on all CDMs received up to now, with associated uncertainty estimates about all predictions. Furthermore Kessler includes a novel generative model of conjunction events and CDM sequences implemented using probabilistic programming, simulating the CDM generation process of the Combined Space Operations Center (CSpOC). The model allows Bayesian inference and also the generation of large datasets of realistic synthetic CDMs that we believe will be pivotal to enable further ML approaches given the sensitive nature and public unavailability of real CDM data.

This paper presents a framework for integrating Low-Earth Orbit (LEO) platforms with Non-Terrestrial Networks (NTNs) in the emerging 6G communication landscape. Our work applies the Mega-Constellation Services in Space (MCSS) paradigm, leveraging LEO mega-constellations’ expansive coverage and capacity, designed initially for terrestrial devices, to serve platforms in lower LEO orbits. Results show that this approach overcomes the limitation of sporadic and time-bound satellite communication links, a challenge not fully resolved by available Ground Station Networks and Data Relay Systems.We contribute three key elements: ( i ) a detailed MCSS evaluation framework employing Monte Carlo simulations to assess space user links and distributions; ( ii ) a novel Space User Terminal (SUT) design optimized for MCSS, using different configurations and 5G New Radio Adaptive Coding and Modulation; ( iii ) extensive results demonstrating MCSS’s substantial improvement over existing Ground Station Networks and Data Relay Systems, motivating its role in the upcoming 6G NTNs. The space terminal, incorporating a multi-system, multi-orbit, and software-defined architecture, can handle Terabit-scale daily data volumes and minute-scale latencies. It offers a compact, power-efficient solution for properly integrating LEO platforms as space internet nodes.

The planning and execution of modern space missions rely on traditional SSA methods for detecting and tracking orbiting hazards. This often leads to sub-optimal responses due to remote sensing inaccuracies and transmission delays. On the other hand, deploying and maintaining space-based sensors is expensive and technically challenging due to the inadequacy of current vision technologies. In this paper, we propose a novel perception framework to enhance in-orbit autonomy and address the shortcomings of traditional SSA methods. We leverage the advances of neuromorphic cameras for a vastly superior sensing performance under space conditions. Additionally , we maximize the advantageous characteristics of the sensor by harnessing the modelling power and efficient design of selective State Space Models. Specifically, we introduce two novel event-based backbones, E-Mamba and E-Vim, for real-time on-board inference with linear scaling in complexity w.r.t. input length. Extensive evaluation across multiple neuromorphic datasets demonstrate the superior parameter efficiency or our approaches (

In this work, we propose a design space exploration workflow and tool for generating reconfigurable deep learning hardware models for FPGAs. The workflow is broken down into two main parts, Offline Design Exploration (ODE) and Online Design Reconfiguration (ODR). Offline Design Exploration is automated through a workflow methodology which makes it possible for a designer to provide a Convolutional Neural Network (CNN) architecture and the option of providing additional design constraints in terms of latency and space. These automatically generate multiple design spaces which trade-off latency for resource utilization by dedicating varying processing elements through loop reordering, unrolling and pipelining. The second part of the process introduces online reconfiguration to the design space. Online Reconfiguration means the ability to modify the design at runtime after upload, by selectively running it partially or fully according to an application’s immediate requirements, this provides the design with flexibility which, we believe, is highly beneficial for future autonomous on-board applications. We validate our work on the Xilinx Zynq-7100 board at 200 MHz and use custom trained networks architectures on three datasets for image classification, MNIST, SVHN and CIFAR-10. ODE generated designs achieved latency trade-offs of 95x for MNIST, 71x for CIFAR-10 and 18x for SVHN. Trade-offs in resource utilization in terms of DSP Slices were 44x for MNIST, 52x for SVHN and 24x for CIFAR-10. For the ODR, a 0.7% accuracy loss was traded-off with x13 speedup and a 25% reduction in power for MNIST, a 2% accuracy loss was traded-off with a 14x speedup and a 28% power reduction for SVHN, a 4% accuracy loss was traded off for a 50x speedup with 32.5% power reduction for CIFAR-10.

Developments in technologies, attitudes and investment are transforming the space environment, achieving greater accessibility for an increasing number of parties. New and proposed constellations will increase the in-orbit satellite population by the order of thousands, expanding the threat landscape of the space industry. This article analyses past satellite security threats and incidents to assess the motivations and characteristics of adversarial threats to satellites. The ground and radio frequency communications were the most favoured targets; however, the boom of satellites constellations in the upcoming years may shift this focus towards the space segment which must be addressed. Key technology advancements and open issues in the satellite industry related to security and operational requirements are also discussed.

As megaconstellations are launched and the space sector grows, space debris pollution is posing an increasing threat to operational spacecraft. Low Earth orbit is a junkyard of dead satellites, rocket bodies, shrapnels, and other debris that travel at very high speed in an uncontrolled manner. Collisions at orbital speeds can generate fragments and potentially trigger a cascade of more collisions endangering the whole population, a scenario known since the late 1970s as the Kessler syndrome. In this work we present Kessler: an open-source Python package for machine learning (ML) applied to collision avoidance. Kessler provides functionalities to import and export conjunction data messages (CDMs) in their standard format and predict the evolution of conjunction events based on explainable ML models. In Kessler we provide Bayesian recurrent neural networks that can be trained with existing collections of CDM data and then deployed in order to predict the contents of future CDMs in a given conjunction event, conditioned on all CDMs received up to now, with associated uncertainty estimates about all predictions. Furthermore Kessler includes a novel generative model of conjunction events and CDM sequences implemented using probabilistic programming, simulating the CDM generation process of the Combined Space Operations Center (CSpOC). The model allows Bayesian inference and also the generation of large datasets of realistic synthetic CDMs that we believe will be pivotal to enable further ML approaches given the sensitive nature and public unavailability of real CDM data.

The InflateSail (QB50-UK06) CubeSat, designed and built at the Surrey Space Centre (SSC) for the Von Karman Institute (VKI), Belgium, was a technology demonstrator built under the European Commission’s QB50 programme. The 3.2 kilogram 3U CubeSat was equipped with a 1 metre long inflatable mast and a 10m2 deployable drag sail and was one of 31 satellites that were launched simultaneously on the PSLV (polar satellite launch vehicle) C-38 from Sriharikota, India on 23rd June 2017 into a 505km, 97.44o Sun-synchronous orbit. Shortly after insertion into orbit, InflateSail automatically activated its drag-sail payload, and, as planned, began to lose altitude, causing it to re-enter the atmosphere just 72 days later – successfully demonstrating for the first time the de-orbiting of a spacecraft using European inflatable and drag-sail technologies. This paper discusses the dynamics we observed during the descent, including the sensitivity of the craft to atmospheric density changes. The InflateSail project was funded by two European Commission Framework Program Seven (FP7) projects: DEPLOYTECH and QB50. QB50 was a programme, led by VKI, for launching a network of 50 CubeSats built mainly by university teams all over the world to perform first-class science in the largely unexplored lower thermosphere.

With the space shuttle on the eve of its final mission, British companies are at the forefront of innovation to drive the next wave of space exploration

Applications such as disaster management enormously benefit from rapid availability of satellite observations. Traditionally, data analysis is performed on the ground after being transferred-downlinked-to a ground station. Constraints on the downlink capabilities, both in terms of data volume and timing, therefore heavily affect the response delay of any downstream application. In this paper, we introduce RaV ae n, a lightweight, unsupervised approach for change detection in satellite data based on Variational Auto-Encoders (VAEs), with the specific purpose of on-board deployment. RaV ae n pre-processes the sampled data directly on the satellite and flags changed areas to prioritise for downlink, shortening the response time. We verified the efficacy of our system on a dataset-which we release alongside this publication-composed of time series containing a catastrophic event, demonstrating that RaV ae n outperforms pixel-wise baselines. Finally, we tested our approach on resource-limited hardware for assessing computational and memory limitations, simulating deployment on real hardware.

The use of commercial of the shelf (COTS) processors is increasingly attractive for the space domain, especially with emerging high demand applications in Earth observation and communications. An order of magnitude improvement in on-board processing capability with less size, mass, and power is possible, however, COTS parts still lag in terms of reliability in the space environment. Costly protection techniques to ensure resilience to single event effects (SEEs) is required. Whilst current software reliability techniques are only capable of detecting errors, and performing partial recovery, our research offers a step change for both error detection and recovery without degradation in fault coverage. This targets modern multicore processors. We have previously shown how to create additional passes in the compiler's intermediate representation layer to automatically add differing protection codes at compile-time using the LLVM compiler framework. LLVM is supported by multiple processing architectures, and multiple high level languages - meaning it can be ported to not just space applications, but aerospace, defence, medical, and automotive. In this paper a new LLVM fault injection tool is presented to validate and measure software protection methods - either statically at compile time or dynamically at runtime for multiple errors such as silent data corruption (SDC), control/flow errors, and crashes. We use our tool to inject faults into unprotected and protected codes and make quantitative comparisons of the errors and associated statistical confidence. Our protection method shows high coverage, up to 100% for some benchmarks, and does not assume that the memory system is protected via typical TMR hardware approaches. This means that we protect all memory instructions that use read and write. Another reason for the high coverage is the inclusion of multiple data and instruction types (i32, i32*, i1, i8, i8*, i64, float & double, float & double pointers). This research has been implemented in two processing architectures; Intel core i5-3470 with 3.2 GHz frequency and a Raspberry Pi 3. On the 1 st processing platform the overhead was less than 15% and on the 2 nd platform the overhead was less than 17%.

Significant advances in spaceborne imaging payloads have resulted in new big data problems in the Earth Observation (EO) field. These challenges are compounded onboard satellites due to a lack of equivalent advancement in onboard data processing and downlink technologies. We have previously proposed a new GPU accelerated onboard data processing architecture and developed parallelised image processing software to demonstrate the achievable data processing throughput and compression performance. However, the environmental characteristics are distinctly different to those on Earth, such as available power and the probability of adverse single event radiation effects. In this paper, we analyse new performance results for a low power embedded GPU platform, investigate the error resilience of our GPU image processing application and offer two new error resilient versions of the application. We utilise software based error injection testing to evaluate data corruption and functional interrupts. These results inform the new error resilient methods that also leverages GPU characteristics to minimise time and memory overheads. The key results show that our targeted redundancy techniques reduce the data corruption from a probability of up to 46% to now less than 2% for all test cases, with a typical execution time overhead of 130%.

Multicellular system architectures are proposed as a solution to the problem of low reliability currently seen amongst small, low cost satellites. In a multicellular architecture, a set of independent k-out-of-n systems mimic the cells of a biological organism. In order to be beneficial, a multicellular architecture must provide more reliability per unit of overhead than traditional forms of redundancy. The overheads include power consumption, volume and mass. This paper describes the derivation of an analytical model for predicting a multicellular system's lifetime. The performance of such architectures is compared against that of several common forms of redundancy and proven to be beneficial under certain circumstances. In addition, the problem of peripheral interfaces and cross-strapping is investigated using a purpose-developed, multicellular simulation environment. Finally, two case studies are presented based on a prototype cell implementation, which demonstrate the feasibility of the proposed architecture.

The AlSat-Nano mission is a joint endeavour by the UK and Algeria to build and operate a 3U CubeSat. The project was designed to provide training to Algerian students, making use of UK engineering and experience. The CubeSat was designed and built by the Surrey Space Centre (SSC) of the University of Surrey and hosts three UK payloads with operations run by the Algerian Space Agency (ASAL). The educational and CubeSat development were funded by the UK Space Agency (UKSA), whilst the UK payloads were self-funded. Launch and operations are funded by ASAL. This paper illustrates the development of the programme, the engineering of the satellite and the development of collaborative operations between the SSC and ASAL.

Over 34,000 objects bigger than 10 cm in length are known to orbit Earth. Among them, only a small percentage are active satellites, while the rest of the population is made of dead satellites, rocket bodies, and debris that pose a collision threat to operational spacecraft. Furthermore, the predicted growth of the space sector and the planned launch of megaconstellations will add even more complexity, therefore causing the collision risk and the burden on space operators to increase. Managing this complex framework with internationally agreed methods is pivotal and urgent. In this context, we build a novel physics-based probabilistic generative model for synthetically generating conjunction data messages, calibrated using real data. By conditioning on observations, we use the model to obtain posterior distributions via Bayesian inference. We show that the probabilistic programming approach to conjunction assessment can help in making predictions and in finding the parameters that explain the observed data in conjunction data messages, thus shedding more light on key variables and orbital characteristics that more likely lead to conjunction events. Moreover, our technique enables the generation of physically accurate synthetic datasets of collisions, answering a fundamental need of the space and machine learning communities working in this area. Third Workshop on Machine Learning and the Physical Sciences (NeurIPS 2020), Vancouver, Canada.

Today's mobile devices and countless other embedded devices now aim to use networking technologies utilizing the latest electronics and software to provide new functions. Distributed satellite systems, seen to be analogous to mobile ad hoc networks (MANET), perform new mission functions with high mobility and intermittent connectivity that make satellite network management and operations difficult. New drivers and requirements are outlined for node and network levels in any given topology requiring real-time client-server or peer-to-peer (P2P) networking applications. To meet these requirements a novel agent computing platform (ACP) is proposed utilizing technologies from the multi-processor and agent middleware fields for real-time Java networking and mobile ad hoc network-based distributed computing applications at a minimal overhead to existing systems. The Java optimised processor (JOP) is investigated and embedded into an existing LEON3-based system-on-a-chip (SoC) design to provide a new fault-tolerant, parallel processing, and network functionalities. Agent middleware is discussed and compared for porting to the new dual processor design with a new middleware instance manager thread to enable software resets at runtime on the Java processor without halting the processor. After verification these two technologies are combined and discussed in depth to highlight key technological problems of this real-time ACP implementation.

Continual advancements in Earth Observation (EO) optical imager payloads has led to a significant increase in the volume of multispectral data generated onboard EO satellites. As a result, a growing onboard data bottleneck need to be alleviated. One technique commonly used is onboard image compression. However, the performance of traditional space qualified processors, such as radiation hardened FPGAs, are not able to meet current nor future onboard data processing requirements. Therefore, a new high capability hardware architecture is required. In previous work a new GPU accelerated scalable heterogeneous hardware architecture for onboard data processing was proposed. In this paper, two new CUDA GPU implementations of the state-of-the-art lossless multidimensional image compression algorithm CCSDS-123, are discussed. The first implementation is a generic CUDA implementation of the CCSDS-123 algorithm whilst the second is optimised specifically for multispectral EO imagery. Both implementations utilise image tiling to leverage an additional axis for algorithm parallelisation to increase processing throughput. The CUDA implementation and optimisation techniques deployed are discussed in the paper. In addition, compression ratio and throughput performance results are presented for each implementation. Further experimental studies into the relationships between algorithm user definable compression parameters, tile sizes, tile dimensions and the achieved compression ratio and throughput, were performed.

SURO-LC is the radio astronomer's equivalent of the first high resolution X-ray space telescopes. It opens up a largely unexplored spectral band, previously hidden from Earth, to make new discoveries in the nearby and distant universe. The proposed mission offers the first omnidirectional low frequency radio survey at high sensitivity and high resolution. SURO-LC all-sky or rapid monitoring (for rapid solar and galactic events) operation is in the largely un-explored frequency domain between 0.1 and 70 MHz, of which the 0.1 - 30 MHz range is mostly inaccessible from earth because of ionospheric blocking and man-made radio frequency interference (RFI). SURO-LC deploys a formation of nine spacecraft in a low relative-drift Lissajous orbit at SEL2, 1.5 million km from earth in a radio clean environment. Eight spherically distributed Cubesat daughters, equipped with 3 orthogonal dipole antennas, form a distributed interferometric radio telescope. An offset mothership provides data acquisition, digital signal processing, and ground communication.

Today’s mobile devices and countless other embedded devices now aim to use networking technologies utilizing the latest electronics and software to provide new functions. Distributed satellite systems, seen to be analogous to mobile ad-hoc networks, perform new mission functions with high mobility and intermittent connectivity which make satellite network management and operations difficult. New drivers and requirements are outlined for node and network levels in any given topology requiring real-time client-server or peer-to-peer networking applications. To meet these requirements, a novel agent computing platform is proposed utilizing technologies from the multi-processor and agent middleware fields for real-time Java networking and mobile ad-hoc network based distributed computing applications at a minimal overhead to existing systems. The Java Optimised Processor (JOP) is investigated and embedded into an existing LEON3 based system-on-a-chip design to provide a new fault-tolerant, parallel processing, and network functionalities. Agent middleware is discussed and compared for porting to the new dual processor design with a new middleware instance manager thread to enable software resets at runtime on the Java processor without halting the processor. After verification, these two technologies are combined and discussed in depth to highlight key technological problems of this real-time agent computing platform implementation.

Low-cost satellites continue to grow in popularity and capability, but have shown poor on-orbit performance to date. While traditional satellite missions have relied upon expensive fault prevention techniques, such as component screening, the use of radiation hardened components, and extensive test campaigns, low-cost missions must focus on fault tolerance, instead. This paper describes a novel, fault-tolerant system architecture, named Satellite Stem Cells. The Satellite Stem Cell Architecture, which is based on artificial cells, evolved from research into traditional reliability theory, bio-inspired engineering, and agentbased computing. Traditional reliability theory points towards k-out-of-n architectures for their superior reliability, while cell biology demonstrates how to build extremely multifunctional subsystems. Finally, agent computing provides a solution for facilitating the cooperation of a set of autonomous cells in a peer-to-peer environment. This paper describes the development of the architecture, details the artificial cell design, and gives preliminary implementation details

The CubeSat standard has inspired a new wave of engineers, researchers, and scientists – all aiming to utilise „commercial off-the-shelf‟ (COTS) subsystems to build nanosatellite systems. For this same reason, students and staff at Surrey have been designing a 3U CubeSat with the intention of providing low-level design, build and test experience for early career engineers, provide on-orbit demonstration of key technologies in attitude and orbit control systems (AOCS), and finally assess smart-phone components for future spacecraft applications. The modern smart-phone is the latest in a consumer driven and start-of-the-art electronics market which the STRaND-1 mission aims to exploit; assessing hardware components and exploring software towards new methods in designing and operating new satellites. The „Surrey Training, Research and Nanosatellite Demonstrator‟ is the first in a line of ambitious missions which aims to strengthen the ties between Surrey Space Centre (SSC) and Surrey Satellite Technology Ltd (SSTL). The project began in April 2010 where STRaND-1 was designed in the team‟s free time but the major build work began in mid-October 2012. Launched on 25 February 2013, the assembly, integration and test (AIT) phase took approximately 4 months to complete. This paper discusses this phase and the typical-CubeSat subsystems and non-PC/104 AOCS and computing payloads in a custom payload bay. In-orbit results of the new nanosatellite power, attitude and orbit control system are presented. With many new capabilities to demonstrate on STRaND-1 when in-orbit, exploiting the latest consumer electronics and software with real training and mission utility was disruptive. We evaluate the ground-based operational processes during commissioning.

This paper describes an agent platform based on the Foundation for Intelligent Physical Systems Abstract Architecture, which, together with a highly fault tolerant, bio-inspired hardware architecture, aims to increase the reliability of future, low-cost satellites. To achieve the stringent operational requirements imposed by the real-time and resource-constrained environment of a satellite, the Hybrid Agent Real-Time Platform (HARP) distinguishes itself from other platforms in three areas. Firstly, the HARP middleware uses discrete processors, instead of virtual machines or interpreters, as its agent execution environment. This has the advantage of reducing the agency memory footprint and enabling agents to perform real-time tasks. Secondly, the HARP communication stack makes use of ISO-TP over CAN 2.0A as its transfer level protocol, cutting out resource-intensive layers such as HTTP and IIOP. In addition, the communication stack allows real-time CAN traffic to share the network and be given priority over Agent Communication Language messages. Finally, the HARP middleware embeds a peer-to-peer task manager in each agency, allowing systems which are built using the bio-inspired Artificial Stem Cell Architecture and HARP middleware to autonomously reconfigure in the event of failures. The detailed design of the HARP middleware is given, together with details of an implementation of the HARP middleware on a set of prototype satellite hardware. The performance and scaling potential of the middleware, determined through a set of physical experiments, provide evidence of the practical feasibility of the proposed architecture.

Advancements in onboard data processing capabilities of small EO satellites represent an avenue for mission integrators, satellite customers and end-users alike to maximise the return on investment of space-borne remote sensing platforms. Surrey Satellite Technology Limited's (SSTL's) Flexible and Intelligent Payload Chain (FIPC) subsystem is an integrated solution which aims to address the data bottleneck challenges of small EO satellites, leveraging capabilities which include onboard data processing. This publication describes SSTL's recent coupled developments in the FIPC space segment, towards a tightly integrated hardware architecture; a new Linux-based custom onboard processing environment; and an end-user segment with a tailored Application Development Framework. Together these facilitate the deployment of in-house and third-party developed software onboard processing Applications and pipelines, including those which exploit machine learning (ML) libraries and frameworks.

After decades of space travel, low Earth orbit is a junkyard of discarded rocket bodies, dead satellites, and millions of pieces of debris from collisions and explosions. Objects in high enough altitudes do not re-enter and burn up in the atmosphere, but stay in orbit around Earth for a long time. With a speed of 28,000 km/h, collisions in these orbits can generate fragments and potentially trigger a cascade of more collisions known as the Kessler syndrome. This could pose a planetary challenge, because the phenomenon could escalate to the point of hindering future space operations and damaging satellite infrastructure critical for space and Earth science applications. As commercial entities place mega-constellations of satellites in orbit, the burden on operators conducting collision avoidance manoeuvres will increase. For this reason, development of automated tools that predict potential collision events (conjunctions) is critical. We introduce a Bayesian deep learning approach to this problem, and develop recurrent neural network architectures (LSTMs) that work with time series of conjunction data messages (CDMs), a standard data format used by the space community. We show that our method can be used to model all CDM features simultaneously, including the time of arrival of future CDMs, providing predictions of conjunction event evolution with associated uncertainties.

The European Student Earth Orbiter (ESEO) is a micro-satellite mission to Low Earth Orbit and is being developed, integrated, and tested by European university students as an ESA Education Office project. AMSAT-UK and Surrey Space Centre are contributing to the mission with a transceiver and transponder similar to that of FUNcube-1 with the addition of utilising an Atmel AT32 processor for packet software-redundancy, baseband processing, forward error correction, and packet forming; acting as a step towards software defined radio using automotive microprocessors [1]. As on the FUNcube-1 satellite, the telemetry formats and encoding schemes presented utilize a large ground network of receivers on the VHF downlink and conforms to 1200 bps and a new 4800 bps redundant downlink for the rest of the spacecraft. The uplink is on L-band using bespoke partial-CCSDS frames. This paper describes the lean satellite design approach introduced by Cho et al. [2] for hardware and software development and testing of the proto-flight model (PFM) payload computer. Furthermore, it assesses the compliance of the project to customer and ESA specifications and discusses the applicability of these standards. Finally, lessons learned are elaborated to provide guidance for future small satellite projects. Through multiple student projects, it was possible to successfully develop a proto-flight model using the lean satellite design approach which entailed an improvement of customer specification compliance from 81% to 86% comparing to the engineering model. In software, utilising the Google Test Suite for verification of the SDR functions and FreeRTOS tools allowed students to optimize processor load margins to 30% when operating parallelized ADC and DAC, and CAN-open telemetry chains and exploring stable memory operations. A further finding was that in Summer 2017, there was an overall compliance of 82% to the CubeSat standard and 57% to the analysed set of ECSS specifications could be achieved. The poorer compliance in ECSS is due to the incomplete environmental testing at that time. The unfunded and student-based nature of the project places significant challenges when compared to conventional missions – but this was outweighed by the ESEO flight opportunity. Following this, we recommended to further the development of a new ISO standard for lean satellite design as initiated by Cho et al. [3] which eases the development process and reliability of small space projects that struggle to fully comply to ECSS or CubeSat specifications. ESA have since defined a subset of ECSS Specifications for educational and CubeSat missions.

The Surrey Training Research and Nanosatellite Demonstrator (STRaND) programme has been success in identifying and creating a leading low-cost nanosatellite programme with advanced attitude and orbit control system (AOCS) and experimental computing platforms based on smart-phone technologies. The next demonstration capabilities, that provide a challenging mission to the existing STRaND platform, is to perform visual inspection, proximity operations and nanosatellite docking. Visual inspection is to be performed using a COTS LIDAR system to estimate range and pose under 100 m. Proximity operations are controlled using a comprehensive guidance, navigation and control (GNC) loop in a polar form of the Hills Clohessy Wiltshire (HCW) frame including J2 perturbations. And finally, nanosatellite docking is performed at under 30 cm using a series of tuned magnetic coils. This paper will document the initial experiments and calculations used to qualify LIDAR components, size the mission thrust and tank requirements, and air cushion table demonstrations of the docking mechanism.

The InflateSail (QB50-UK06) CubeSat, designed and built at the Surrey Space Centre (SSC) for the Von Karman Institute (VKI), Belgium, was one of the technology demonstrators for the European Commission’s QB50 programme. The 3.2 kg 3U CubeSat was equipped with a 1 metre long inflatable mast and a 10m2 deployable drag sail. InflateSail's primary mission was to demonstrate the effectiveness of using a drag sail in Low Earth Orbit (LEO) to dramatically increase the rate at which satellites lose altitude and re-enter the Earth's atmosphere and it was one of 31 satellites that were launched simultaneously on the PSLV (polar satellite launch vehicle) C-38 from Sriharikota, India on 23rd June 2017 into a 505km, 97.44o Sun-synchronous orbit. Shortly after safe deployment in orbit, InflateSail automatically activated its payload. Firstly, it inflated its metrelong metal-polymer laminate tubular mast, and then activated a stepper motor to extend four lightweight bi-stable rigid composite (BRC) booms from the end of the mast, so as to draw out the 3.1m x 3.1m square, 12m thick polyethylene naphthalate (PEN) drag-sail. As intended, the satellite immediately began to lose altitude, causing it to re-enter the atmosphere just 72 days later – thus successfully demonstrating for the first time the de-orbiting of a spacecraft using European inflatable and drag-sail technologies. The InflateSail project was funded by two European Commission Framework Program Seven (FP7) projects: DEPLOYTECH and QB50. DEPLOYTECH had eight European partners including DLR, Airbus France, RolaTube, Cambridge University, and was assisted by NASA Marshall Space Flight Center. DEPLOYTECH’s objectives were to advance the technological capabilities of three different space deployable technologies by qualifying their concepts for space use. QB50 was a programme, led by VKI, for launching a network of 50 CubeSats built mainly by university teams all over the world to perform first-class science in the largely unexplored lower thermosphere. The boom/drag-sail technology developed by SSC will next be used on a third FP7 Project: RemoveDebris, launched in 2018, which will demonstrate the capturing and de-orbiting of artificial space debris targets using a net and harpoon system. This paper describes the results of the InflateSail mission, including the observed effects of atmospheric density and solar activity on its trajectory and body dynamics. It also describes the application of the technology to RemoveDebris and its potential as a commercial de-orbiting add-on package for future space missions.

A new class of low-cost satellites has the potential to reduce the cost of traditional space-based services. Unfortunately, to date, low-cost satellites have proven to suffer from poor reliability. While traditional techniques for increasing reliability are well known to satellite developers, these techniques are poorly suited for implementation on low-cost satellites due to intrinsic budgetary, mass and volume constraints. This research proposes that alternative techniques for increasing system reliability can be derived by studying biological organisms, which have proven their robustness by inhabiting even the harshest locations on earth. Both unicellular and multicellular organisms are examined. The result is a conceptual system architecture, based on initially identical, reconfigurable hardware blocks, or artificial cells, and a decentralized task management strategy. This multicellular architecture is described in detail. Finally, preliminary details of a planned implementation are given. This implementation aims to demonstrate that a significant portion of traditional satellite avionics can be replaced by the proposed artificial cells.

© 2015 IEEE.Software Defined Radio (SDR) is a key area to realise new software implementations for adaptive and reconfigurable communication systems without changing any hardware device or feature. A review on efficient use of limited bandwidth and increasing distributed satellite missions can lead to the need for a generic yet configurable communication platform that can handle multiple signals from multiple satellites with various modulation techniques, data rates and frequency bands that must be compatible to typical small satellite requirements. SDR is beneficial for space applications as it can provide the flexibility and re-configurability and this is driven by fast development times, new found heritage, reduced cost, and low mass Commercial Off-The-Shelf (COTS) components. The implementation of a combined System-On-Chip (SoC) and SDR communication platform enables additional reduction in cost as well as mass. This paper proposes a SDR architecture in which Field Programmable Gate Array (FPGA) System-on-Chip (SoC) is paired with a Radio Frequency (RF) programmable transceiver SoC to solve back-end and front-end re-configurability challenges respectively. The test-bed is aimed at implementing the signal processing software functions in both the dual-core ARM processors and associated FPGA fabric. The distribution of the functions between the FPGA fabric and dual-processor is based on profiling experiments using signal processing blocks, implemented on the development platform, in order to identify where bottlenecks exist. This paper discusses further the results from the new multi-signal / multi-satellite pipeline architecture and the subsequent bandwidth, data rate and processing requirements. Aspects of implementing and testing signal processing chains needed for CubeSat Telecommand, Telemetry and Control (TT&C) are presented together with initial results. Thus the proposed technology not only contributes for a lightweight and portable ground station but also for an on-board satellite transceiver.

This paper is concerned with application of standard wireless COTS protocols to space. Suitability of commercially available wireless sensor mote kits for communication inside and between satellites is investigated Spacecraft applications of motes are being considered and a set of requirements are identified Selected mote kits are tested under various scenarios complying with spacecraft testing procedures. The paper details the results of the carried out functional, EMC/I, vibration, thermal and radiation tests.

This paper presents the results of a research project, which aims to develop enabling technologies for future distributed space architectures based on flexible, reconfigurable, evolvable, and intelligent multi-spacecraft sensing networks. One important goal of the project is to propose a distributed computing platform over wireless inter-satellite links. The paper discusses initial results on the application of distributed computing technologies to future networked constellations of picosatellites.

Understanding the lunar near-surface distribution of relevant in-situ resources, such as ilmenite (FeTiO3), and volatiles, such as water/ice, is vital to future sustained manned bases. VMMO is a highly-capable, low-cost 12U Cubesat designed for operation in a lunar frozen orbit. It accomodates the LVMM Lunar Volatile and Mineralogy Mapper and the CLAIRE Compact LunAr Ionising Radiation Environment payloads. LVMM is a multi-wavelength Chemical Lidar using fiber lasers emitting at 532nm and 1560nm, with an optional 1064nm channel, for stand-off mapping of the lunar ice distribution using active laser illumination, with a focus on the permanently-shadowed craters in the lunar south pole. This combination of spectral channels can provide sensitive discrimination of water/ice in various regolith. The fiber-laser technology has heritage in the ongoing Fiber Sensor Demonstrator flying on ESA's Proba-2. LVMM can also be used in a low-power passive mode with an added 280nm UV channel to map the lunar mineralogy and ilmenite distribution during the lunar day using the reflected solar illumination. CLAIRE is designed to provide a highly miniaturized radiation environment and effect monitor. CLAIRE draws on heritage from the MuREM and RM payloads, flown on the UK’s TDS-1 spacecraft. The payload includes PIN-diode sensors to measure ionizing particle fluxes (protons and heavy-ions) and to record the resulting linear energy transfer (LET) energy-deposition spectra. It also includes solid-state RADFET dosimeters to measure accumulated ionizing dose, and dose-rate diode detectors, designed to respond to a Coronal Mass Ejection (CME) or Solar Particle Event (SPE). CLAIRE also includes an electronic component test board, capable of measuring SEEs and TID effects in a selected set of candidate electronics, allowing direct correlations between effects and the real measured environment.

Future space telescopes with diameter over 20 m will require in-space assembly. High-precision formation flying has very high cost and may not be able to maintain stable alignment over long periods of time. We believe autonomous assembly is a key enabler for a lower cost approach to large space telescopes. To gain experience, and to provide risk reduction, we propose a demonstration mission to demonstrate all key aspects of autonomous assembly and reconfiguration of a space telescope based on multiple mirror elements. The mission will involve two 3U CubeSat-like nanosatellites (“MirrorSats”) each carrying an electrically actuated adaptive mirror, and each capable of autonomous un-docking and re-docking with a small central “9U” class nanosatellite core, which houses two fixed mirrors and a boom-deployed focal plane assembly. All three spacecraft will be launched as a single ~40kg microsatellite package.

The European Student Earth Orbiter (ESEO) is a micro-satellite mission to Low Earth Orbit and is being developed, integrated, and tested by European university students as an ESA Education Office project. AMSAT-UK and Surrey Space Centre are contributing to the mission with a transceiver and transponder similar to that of FUNcube-1 with the addition of utilising a Atmel AT32 processor for packet software-redundancy, baseband processing, forward error correction, and packet forming; acting as a step towards software defined radio using low MIPS automotive microprocessors. As on the FUNcube-1 satellite, the telemetry formats and encoding schemes presented utilize a large ground network of receivers on the VHF downlink and conforms to 1200 bps and a new 4800 bps redundant downlink for the rest of the spacecraft. The uplink is on L-band using bespoke partial-CCSDS frames. This paper details the flight software on the engineering and flight models to ESA, and the technical configuration and associated tests of demonstrating the processor load is under for varying operating and sampling modes. In particular, a key contribution will be the details of utilising the Google Test Suite for verification of the SDR functions and FreeRTOS tools to optimize processor load margins to 30% when operating parallelized ADC and DAC, and CAN-open telemetry chains and what memory considerations are needed to ensure stable long-term operations.

Surrey Satellite Technology Ltd and the University of Surrey have a long history of demonstrating new terrestrial COTS technologies in space, with the aim of reducing the cost of space applications. The STRaND-1 mission is the most recent example of this cooperative history, demonstrating the use of mobile phone technology as the central avionics for a nanosatellite. The STRaND-1 mission is planned for launch in 2012. The second mission in the STRaND programme is now under development, under the same funding arrangement as STRaND-1 (equally and internally funded by SSTL and SSC), and the aim is to be just as ambitious as the first STRaND mission. One key technology to demonstrate is the use of a Microsoft Kinect (TM) sensor suite in orbit, as a low cost alternative to lidar and machine imaging, to enable a docking mission between two CubeSats. If successful, the STRaND-2 mission would be the first demonstration of autonomous docking of nano-scale spacecraft. This paper discusses how the STRaND philosophy can be applied to a second mission. It then outlines the mission concept, highlighting key technology areas including the docking suite and propulsion system, before detailing some of the orbit dynamics of docking two CubeSats together. Initial mass and power budgets are provided with an overview of the system design. The paper concludes with a discussion on the future applications enabled by CubeSat docking technologies, including the AAReST mission concept - a collaborative mission between the University of Surrey and CalTech and NASA JPL. Copyright © (2012) by the International Astronautical Federation.

AMSAT-UK and the Surrey Space Centre are cooperating in delivering an educational communication payload for the ESA European Student Earth Orbiter (ESEO) mission, comprising a payload computer, an L-band receiver and a VHF transmitter. The primary purpose of the payload is to provide downlink telemetry that can be easily received by schools and colleges for educational outreach purposes [1]. Common space industry standards such as European Cooperation for Space Standardization (ECSS) consist of a large number of documents that were primarily written for large-scale space missions. Academic space projects cannot follow these design guidelines due to a lack of sufficient expertise, human resources, facilities or equipment. However, many projects were successfully developed, launched and operated with major deviations from ECSS standards. A recently published CubeSat standard consists of tailored ECSS requirements with the aim to improve the applicability of these specifications to small satellite projects. These, however, are still incompatible with the limited working environment of most university projects. In recent years, a `lean satellite' design approach that utilises non-traditional, risk-taking development and management was proposed by Cho et al. [2] to address these issues. This design approach was successfully applied by the AMSAT project team to develop a proto-flight model of the payload which entailed an improvement of customer specification compliance from 81% to 86% with respect to the engineering model. This method allowed a low cost and fast development process as well as passing all functional and environmental tests without major issues. A key finding was that despite having superior facilities, equipment and expertise compared to most academic CubeSat teams, only an overall compliance of 82% to the CubeSat standard and 57% to the analysed set of ECSS specifications could be achieved. This shows the challenge small space projects face when following conventional industry specifications such as ECSS which are written for traditional space missions. Following this, it is recommended to further promote the development of a new ISO standard for lean satellite design which could ease the development process and reliability of small space projects that struggle to fully comply to ECSS or CubeSat specifications.

The European Student Earth Orbiter (ESEO) is a micro-satellite mission to Low Earth Orbit and is being developed, integrated, and tested by European university students as an ESA Education Office project. AMSAT-UK and Surrey Space Centre are contributing to the mission with a transceiver and transponder similar to that of FUNcube-1 with the addition of utilising an Atmel AT32 processor for packet software-redundancy, baseband processing, forward error correction, and packet forming; acting as a step towards software defined radio using automotive microprocessors [1]. As on the FUNcube-1 satellite, the telemetry formats and encoding schemes presented utilize a large ground network of receivers on the VHF downlink and conforms to 1200 bps and a new 4800 bps redundant downlink for the rest of the spacecraft. The uplink is on L-band using bespoke partial-CCSDS frames. This paper describes the lean satellite design approach introduced by Cho et al. [2] for hardware and software development and testing of the proto-flight model (PFM) payload computer. Furthermore, it assesses the compliance of the project to customer and ESA specifications and discusses the applicability of these standards. Finally, lessons learned are elaborated to provide guidance for future small satellite projects. Through multiple student projects, it was possible to successfully develop a proto-flight model using the lean satellite design approach which entailed an improvement of customer specification compliance from 81% to 86% comparing to the engineering model. In software, utilising the Google Test Suite for verification of the SDR functions and FreeRTOS tools allowed students to optimize processor load margins to 30% when operating parallelized ADC and DAC, and CAN-open telemetry chains and exploring stable memory operations. A further finding was that in Summer 2017, there was an overall compliance of 82% to the CubeSat standard and 57% to the analysed set of ECSS specifications could be achieved. The poorer compliance in ECSS is due to the incomplete environmental testing at that time. The unfunded and student-based nature of the project places significant challenges when compared to conventional missions – but this was outweighed by the ESEO flight opportunity. Following this, we recommended to further the development of a new ISO standard for lean satellite design as initiated by Cho et al. [3] which eases the development process and reliability of small space projects that struggle to fully comply to ECSS or CubeSat specifications. ESA have since defined a subset of ECSS Specifications for educational and CubeSat missions.

The advent of mega-constellations has given rise to an unprecedented exponential growth in the numbers of objects in orbit. As the number of objects sharing similar orbital trajectories increases, as do the probabilities of close encounters and subsequent collisions. Collisions produce more objects further increasing the probability of later collisions until the Earth orbit environment is rendered unusable. Accurate prediction of these encounters is key to enabling satellite operators to perform collision avoidance manoeuvres. This prediction is typically performed by a chain of 1) Orbital Propagators, to determine an objects state vector at a given time 2) Encounter Analysers, to determine which objects are sufficiently close to warrant further examination and 3) Statistical Models, to determine the probability that a conjunction will result in a collision. There is a need to rapidly compute data for a timelier response to threats as new observation data is received. The generation of both historical & augmented-future observation data can improve existing statistical models or as training data for machine learning systems. A key issue is how to ingest, propagate and provide statistics on historical observation data from CSpOC – estimated to take >20 years on a naive CPU only pipeline, and >6 months on a traditionally optimised solution.

The InflateSail (QB50-UK06) CubeSat, designed and built at the Surrey Space Centre (SSC) for the Von Karman Institute (VKI), Belgium, was one of the technology demonstrators for the European Commission's QB50 programme. The 3.2 kg 3U CubeSat was equipped with a 1 m long inflatable mast and a 10 m2 deployable drag sail. InflateSail's primary mission was to demonstrate the effectiveness of using a drag sail in Low Earth Orbit (LEO) to dramatically increase the rate at which satellites lose altitude and re-enter the Earth's atmosphere and it was one of 31 satellites that were launched simultaneously on the PSLV (polar satellite launch vehicle) C-38 from Sriharikota, India on 23rd June 2017 into a 505 km, 97.44° Sun-synchronous orbit. Shortly after safe deployment in orbit, InflateSail automatically activated its payload. Firstly, it inflated its metre-long metal-polymer laminate tubular mast, and then activated a stepper motor to extend four lightweight bi-stable rigid composite (BRC) booms from the end of the mast, so as to draw out the 3.1 m × 3.1 m square, 12 μm thick polyethylene naphthalate (PEN) drag-sail. As intended, the satellite immediately began to lose altitude, causing it to re-enter the atmosphere just 72 days later – thus successfully demonstrating for the first time the de-orbiting of a spacecraft using European inflatable and drag-sail technologies. The InflateSail project was funded by two European Commission Framework Program Seven (FP7) projects: DEPLOYTECH and QB50. DEPLOYTECH had eight European partners including DLR, Airbus France, RolaTube, Cambridge University, and was assisted by NASA Marshall Space Flight Centre. DEPLOYTECH's objectives were to advance the technological capabilities of three different space deployable technologies by qualifying their concepts for space use. QB50 was a programme, led by VKI, for launching a network of 50 CubeSats built mainly by university teams all over the world to perform first-class science in the largely unexplored lower thermosphere. The mast/drag-sail technology developed by SSC will next be used on a third FP7 Project: RemoveDebris, launched in 2018, which will demonstrate the capturing and de-orbiting of artificial space debris targets using a net and harpoon system. This paper describes the results of the InflateSail mission, including the observed effects of atmospheric density and solar activity on its trajectory and body dynamics. It also describes the application of the technology to RemoveDebris and its potential as a commercial de-orbiting add-on package for future space missions. •Description of the InflateSail QB-50 Spacecraft and mission and its results.•First demonstration of cool gas generator inflated inflatable structures in Europe.•First successful European demonstration of using a drag sail to cause re-entry of a spacecraft.•First detailed observations of orbit and body descent using such technology.•Discussion of future application to tackling the space debris problem.

In recent years the growth in quantity, diversity and capability of Earth Observation (EO) satellites, has enabled increase's in the achievable payload data dimensionality and volume. However, the lack of equivalent advancement in downlink technology has resulted in the development of an onboard data bottleneck. This bottleneck must be alleviated in order for EO satellites to continue to efficiently provide high quality and increasing quantities of payload data. This research explores the selection and implementation of state-of-the-art multidimensional image compression algorithms and proposes a new onboard data processing architecture, to help alleviate the bottleneck and increase the data throughput of the platform. The proposed new system is based upon a backplane architecture to provide scalability with different satellite platform sizes and varying mission's objectives. The heterogeneous nature of the architecture allows benefits of both Field Programmable Gate Array (FPGA) and Graphical Processing Unit (GPU) hardware to be leveraged for maximised data processing throughput.

Space traffic is considered a complex contemporary issue originated from the difficulty inherited in the nature of space missions. Satellites and space objects are actively tracked using Earth-based surveillance networks or insitu sensors like the GPS. The data collected from these Earth-based surveillance networks albeit their public availability, they remain only valuable for trajectory and state verification due to their limited positional accuracy. This is exhibited in the poor accuracy (within 2-5 km) of public two-line elements (TLEs), which worsen when propagated in the future to more than tens of kilometers. Consequently, precision orbit data obtained from onboard GPS systems, are commonly used to alleviate, to a degree, the uncertainty regarding space conjunctions. However, the lack of GPS data standardization and sharing regarding critical satellite conjunctions together with the limited deployment of the GPS system onboard nano-satellites and its susceptibility to Doppler shifts in low Earth orbits, inhibits the full exploitation of the GPS system in space traffic. Targeting instant inter-satellite ranging for improving space situational awareness (SSA), this research explores the benefits of the proposed custom signature sequence to optimize a signal acquisition and detection for coherent and asynchronous synchronisation. This comes with the advantage of accurate propagation delay estimation and bias errors correction. Data sharing also becomes possible using the proposed custom direct sequence code division multiple access (DS-CDMA) and Reed-Solomon forward error correction for a futureproof inter-satellite ranging and data sharing solution. Coherent detection, acquisition and synchronisation are proposed based on data-assisted correlation which reduces the false- and miss-alarms respectively by 27% and 97%. This CDMA frame is based on orthogonal dual-channel separation which offers considerable noise and interference resilience for data recovery, in addition to precise frequency and timing errors correction during synchronisation. Results show that the ranging accuracy approaches meter-level at no expense of increased bandwidth beyond 2 MHz. Occupying the same bandwidth, the ranging resolution scales linearly with the spreading factor contributed by the Kronecker filter length. Additionally, the delay measurement solution is proven unsusceptible to the Doppler frequency errors caused by the high relative velocity between the satellites and Doppler errors within 16 kHz were corrected. Further, the coding gain increases by 16 dB compared to narrow-band communications, using CCSDS Reed-Solomon encoder coupled with data sequence spreading, offering considerable data transfer resilience in low signal-to-noise conditions.

The world space economy is expected to grow to $400 billion by 2030 and to provide 100,000 jobs. In the UK we currently have 38,500 directly employed with a further 70000 jobs dependent on the space sector. By 2030 the UK aims to have a further 100,000 new people employed within the sector. Training space engineers and scientists is critical to fulfilling this need. The UK-based “Space Universities Network” (SUN) was formed in 2016 with the aim of enhancing the quality of learning and teaching by providing support and resources to the Space science and engineering higher education community. SUN’s objectives are to facilitate the creation of a skilled workforce of graduates who can meet the challenges of future scientific and commercial exploitation of space. The network addresses this need by helping to inspire students to join the space sector and ensuring they are well equipped at University to contribute. SUN enables the developing, sharing and promotion of effective practice and innovation in the delivery of university-level space science and engineering curricula. It does this through workshops, offering opportunities for networking to support the space teaching community and a web-based repository of resources. This paper describes the process that led to the foundation of SUN, its objectives, modes of operation, prime activities, evaluation and future projects. Once firmly established, it is hoped to expand the network through partnerships with similar organisations in other countries.

Radio images of red-shifted 21-cm signals from neutral hydrogen originating from the very early Universe, the so-called Dark Ages before the first stars formed, are impossible to obtain from Earth due to man-made radio frequency interference (RFI) and the opacity of the ionosphere below ∼30 MHz. To efficiently block the RFI, which would otherwise overwhelm the weak cosmological signal, requires a large low-frequency radio array on the far-side of the Moon. Such a lander mission is technically challenging and carries a budget that is currently unlikely to be included in any national or international mission plan. Our goal is to use a constellation of small satellites in lunar orbit to collect pathfinder data to demonstrate the feasibility of using the Moon as a radio-shield, and map out the spatial extent of this RF quiescent zone. The team led by the Hawaii Space Flight Laboratory (HSFL) at the University of Hawaii at Manoa is designing a mission to characterize the spatial extent of the RF quiescence zone on the lunar farside to support future missions to explore the cosmos using radio observatories on the surface. This paper examines the design of this mission starting with a baseline architecture that uses a modified SSTL X50 satellite bus as mothership that carries one or more nanosats to lunar orbit. The mothership will then deploy it/them to form the constellation, as well as act as the communications relay between them and Earth. The initial baseline mission utilizes the standard Super Strypi launch vehicle. Although it is desirable to have a mothership and several nanosats evenly distributed in an equatorial lunar orbit, performance limitations of the standard launch vehicle only permit the mothership with one nanosat in a highly elliptical orbit that would allow measurement of the relevant RF environment continuously for at least a year. The nanosat would crosslink the collected data to the mothership, which will relay the data to Earth as well as act as an RF collecting station itself.

The current trend in commercial processors is producing multi-core architectures which pose both an opportunity and a challenge for future space based processing. The opportunity is how to leverage multi-core processors for high intensity computing applications and thus provide an order of magnitude increase in onboard processing capability with less size, mass, and power. The challenge is to provide the requisite safety and reliability in an extremely challenging radiation environment. The objective is to advance from multiple single processor systems typically flown to a fault tolerant multi-core system. Software based methods for multi-core processor fault tolerance to single event effects (SEEs) causing interrupts or ‘bit-flips’ are investigated and we propose to utilize additional cores and memory resources together with newly developed software protection techniques. This work also assesses the optimal trade space between reliability and performance. Our work is based on the modern compiler “LLVM” as it is ported to many architectures, where we implement optimization passes that enable automatic addition of protection techniques including Nmodular redundancy (NMR) and error detection and correction (EDAC) at assembly/instruction level to languages supported. The optimization passes modify the intermediate representation of the source code meaning it could be applied for any high level language, and any processor architecture supported by the LLVM framework. In our initial experiments, we implement separately triple modular redundancy (TMR) and error detection and correction codes including (Hamming, BCH) at instruction level. We combine these two methods for critical applications, where we first TMR our instructions, and then use EDAC as a further measure, when TMR is not able to correct the errors originating from the SEE. Our initial experiments show good performance (about 10% overhead) when protecting the memory of code using double error detection single error correction hamming code and TMR (Triple modular redundancy), further work is needed to improve the performance when protecting the memory of code using the BCH code. This work would be highly valuable, both to satellites/space but also in general computing such as in in aircraft, automotive, server farms, and medical equipment (or anywhere that needs safety critical performance) as hardware gets smaller and more susceptible.

Vast Satcom Antenna (VASANT) is an ongoing system and business case development study for a very large radio-frequency antenna (>>40m diameter) assembled and/or fabricated in space. Such an antenna would be enabling for radical new commercial satcom services to complement the coming broadband satcom constellations. The disruptive capabilities from such a high gain, high bandwidth antenna include potentially building penetration (at VHF) and support for very compact, omni-directional user terminals.

Modern computers are now far in advance of satellite systems and leveraging of these technologies for space applications could lead to cheaper and more capable spacecraft. Together with NASA AMES’s PhoneSat, the STRaND-1 nanosatellite team has been developing and designing new ways to include smart-phone technologies to the popular CubeSat platform whilst mitigating numerous risks. Surrey Space Centre (SSC) and Surrey Satellite Technology Ltd. (SSTL) have led in qualifying state-of-the-art COTS technologies and capabilities - contributing to numerous low cost satellite missions. The focus of this paper is to answer if 1) modern smart-phone software is compatible for fast and low cost development as required by CubeSats, and 2) if the components utilised are robust to the space environment. The STRaND-1 smart-phone payload software explored in this paper is united using various open-source Linux tools and generic interfaces found in terrestrial systems. A major result from our developments is that many existing software and hardware processes are more than sufficient to provide autonomous and operational payload object-to-object and filebased management solutions. The paper will provide methodologies on the software chains and tools used for the STRaND-1 smartphone computing platform, the hardware built with space qualification results (thermal, thermal vacuum, and TID radiation), and how they can be implemented in future missions.

The risk of collisions in Earth’s orbit is growing markedly. In January 2021, SpaceX and OneWeb released an operator-to-operator fact sheet that highlights the critical reliance on conjunction data messages (CDMs) and observations, demonstrating the need for a diverse sensing environment for orbital objects. Recently, the University of Oxford and the University of Surrey developed, in collaboration with Trillium Technologies and the European Space Operations Center, an opensource Python package for modeling the spacecraft collision avoidance process, called Kessler. Such tools can be used for importing/exporting CDMs in their standard format, modeling the current low-Earth orbit (LEO) population and its short-term propagation from a given catalog file, as well as modeling the evolution of conjunction events based on the current population and observation scenarios, hence emulating the CDMs generation process of the Combined Space Operations Center (CSpOC). The model also provides probabilistic programming and ML tools to predict future collision events and to perform Bayesian inference (i.e., optimal use of all available observations). In the framework of a United Kingdom Space Agency-funded project, we analyze and study the impact of megaconstellations and observation models in the collision avoidance process. First, we monitor and report how the estimated collision risk and other quantities at the time of closest approach (e.g. miss distance, uncertainties, etc.) vary, according to different observation models, which emulate different radar observation accuracy. Then, we analyze the impact of future megaconstellations on the number of warnings generated from the increase in the number of conjunctions leading to an increased burden on space operators. FCC licenses were used to identify credible megaconstellation sources to understand how a potential consistent increase in active satellites will impact LEO situational safety. We finally present how our simulations help understand the impact of these future megaconstellations on the current population, and how we can devise better ground observation strategies to quantify future observation needs and reduce the burden on operators.

The increase in satellite launches has led to a jump in the number of satellites orbiting Earth to over 1900 active satellites to date. Most of these satellites rely on the two-line elements (TLEs) to define the satellite tracks. However, the accuracy obtained from TLEs is insufficient for accurate satellite collision predictions which led to the undeniable uncertainty regarding satellite conjunction assessment. This paper extends the research conducted about investigating the implementation of a new inter-satellite ranging instrument by proposing two operational modes namely Discovery and Resolution. Discovery allows long-distance satellite detection and fast range estimation from the received signal strength indicator (RSSI). This ensures a larger observation time-window for the Resolution mode to define precisely the nature of the satellite encounter. The system switches to Resolution when the relative range between the satellites source and observer decreases below 10 km according to the scenario studied. Unlike Discovery, Resolution estimates the range from the round-trip propagation delay using sequential ranging techniques. Results reveal that RSSI range measurements are prone to heavy fluctuations due to the path loss variations. In fact, a standard deviation of 63 km for the ranging errors over 1s measurement time is measured however, these measurements are obtained within 2-mu s time intervals. On the other hand, Resolution measures the range in chips by calculating the argument of the maxima of the cross-correlation function (CCF) output between the transmitted and the received sequences. Results using Resolution show that an accuracy of 110-m is obtained from a ranging sequence of 800 kcps during 1 s observation window. This value is drastically improved compared to the results achieved with Discovery but with the cost of an acquisition-and processing-time of 20 ms compared to 2 mu s attained with Discovery.

Low-cost satellites continue to grow in popularity and capability, but have shown poor on-orbit performance to date. While traditional satellite missions have relied upon expensive fault prevention techniques, such as component screening, the use of radiation hardened components, and extensive test campaigns, low-cost missions must focus on fault tolerance, instead. This paper describes a novel, fault-tolerant system architecture, named Satellite Stem Cells. The Satellite Stem Cell Architecture, which is based on artificial cells, evolved from research into traditional reliability theory, bio-inspired engineering, and agent-based computing. Traditional reliability theory points towards k-out-of-n architectures for their superior reliability, while cell biology demonstrates how to build extremely multifunctional subsystems. Finally, agent computing provides a solution for facilitating the cooperation of a set of autonomous cells in a peer-to-peer environment. This paper describes the development of the architecture, details the artificial cell design, and gives preliminary implementation details.

STRaND-1 is the first in a series of Surrey Satellite Technology Ltd. (SSTL)-Surrey Space Centre (SSC) collaborative satellites designed for the purpose of technology path finding for future commercial operations. It is the first time Surrey has entered the CubeSat field and differs from most CubeSats in that it will fly a modern Commercial Off The Shelf (COTS) Android smartphone as a payload, along with a suite of advanced technologies developed by the University of Surrey, and a payload from the University of Stellenbosch in South Africa. STRaND- 1 is also different in that anyone (not just from the space engineering or space science community) will be eligible to fly their “app" in space, for free. STRaND-1 is currently being manufactured and tested by volunteers in their own free time, and will be ready for an intended launch in the first quarter of 2012. This paper outlines the STRaND pathfinder programme philosophy which challenges some conventional space engineering practises, and describes the impact of those changes on the satellite development lifecycle. The paper then briefly describes the intent behind the design of STRaND-1, before presenting details on the design of the nanosatellite, focussing of the details of the innovative new technologies. These technologies include two different propulsion systems, an 802.11g WiFi experiment, a new VHF/UHF transceiver unit and a miniature 3-axis reaction wheel assembly. The novel processing setup (which includes the smartphone) is discussed in some detail, particularly the potential for outreach via the open source nature of Google's Android operating system. A stepthrough of the planned concept of operations is provided, which includes a possible rendezvous and inspection objective, demonstrating equal or improved capability compared to SNAP-1 with a reduced total system mass. Finally, data from the test campaign is presented and compared against other notable CubeSats known for their advanced capabilities. Rendered images of STRaND-1 are shown in Fig. I and are discussed later in the paper.

This study establishes the optimal Single Event Upset (SEU) mitigation strategy for Xilinx's 7-Series Field Programmable Gate Arrays (FPGAs). This enables 7-Series FPGAs to be utilised in systems with high exposure to ionising radiation over Xilinx's smaller, radiation-hardened 4-Series and 5-Series FPGAs. The optimal strategy maximises system up-time, with minimal complexity and minimal risk of damaging the target FPGA device. Four SEU mitigation techniques are analysed and compared. Three of these are external scrubbing techniques; blind, readback, and blind with Frame Address Register (FAR) verification; with the fourth being Xilinx's internal scrubber, the Soft Error Mitigation (SEM) Intellectual Property (IP) core. The initial comparisons are quantitative, comparing the four methods for traits that lend themselves to meeting the criteria of the optimal scrubber. The techniques are then compared quantitatively to establish theoretical whole-device scan times for each technique given a range of errors to be corrected. These scan times can be utilised to determine which technique performs the fastest given the error rate and technique parameters. The calculations and resultant data can be computed for every device in the 7-series range. The findings suggest that the optimal solution is to opt not for a traditional mitigation strategy, but for a dynamic mitigation strategy. The fastest scrubber is the SEM IP Core, predominantly due to being internal to the FPGA, and as error rates increase, there becomes a crossover point whereby the blind scrubber with FAR verification comes into play to scrub at higher error rates in a deterministic way. Additionally, it finds that the configuration scrubber can operate at higher frequencies in comparison to NASAs recommendation of scrubbing at an order of magnitude higher than the expected SEU rate. These results provide the optimal scrubber for the 7-Series FPGA, and bring into question the recommendation of scrubbing at a higher order of magnitude than the expected error rate. Additionally, Xilinx's guidance for their FPGAs brings in an essential bit ratio which determines what proportion of the FPGAs is actually being utilised by the design, thereby decreasing the proportion of the errors that impact the design as opposed to empty areas of the FPGA, and hence increasing the speed at which scrubbing can take place with reference to NASAs recommendation.

The European Student Earth Orbiter (ESEO) is a micro-satellite mission to low Earth orbit and is being developed, integrated, and tested by European university students as an ESA Education Office project. AMSAT-UK and Surrey Space Centre are contributing to the mission with a transceiver and transponder similar to that of FUNcube-1 with the addition of utilising a Atmel AT32 processor for packet software-redundancy, baseband processing, forward error correction, and packet forming; acting as a step towards software defined radio using low MIPS automotive microprocessors. As on the FUNcube-1 satellite, the telemetry formats and encoding schemes presented utilize a large ground network of receivers on the VHF downlink and conforms to 1200 bps and a new 4800 bps redundant downlink for the rest of the spacecraft. The uplink is on L-band using bespoke partial-CCSDS frames.

A perrenial question of Cubesats is why they are not yet used as platforms for truly operational application missions. The STRaND-1 mission described in this paper is used to demonstrate the hurdles which must be overcome in order to create cost-effective CubeSat platforms that are ready for operational missions with satisfactory design lifetime, reliability and availability objectives. STRaND-1 is the UK’s first CubeSat, and will be launched on the 25th of February 2013 on a PSLV into a dawn-dusk sun-synchronous orbit. As with many CubeSats, the goals of the successful 3U mission were rapid training and technology demonstration. The novelty (other than the technical novelty of testing the robustness of mobile phone electronics in the LEO environment) was the volunteer nature of the team, and that the organisations involved had previous operational small satellite mission experience. This paper takes a holistic view of the mission, critically reviewing the mission lifecycle from the initial concept design through to integration and testing, LEOP and initial mission results in respect of these hurdles to operational applications. The UK's small satellite technology demonstration mission - TDS-1 - is presented for contrast. Now ready for flight in Q3 2013, TDS-1 is an example of how a collaborative small sat technology demonstration mission can be accomplished at low cost and inside a rapid schedule. TDS-1 incorporates a suite of eight separate sensor payloads plus a novel set of advanced avionics. The design, concept of operations and management of TDS-1 enabled the platform enough flexibility to accommodate the payloads to change in both number and in bus resources required throughout the programme lifecycle, while avoiding the pitfalls of over-designing the system. The review is conducted with an eye to how a CubeSat mission differs from the commercial, small satellite approach to spacecraft engineering. In particular, lessons learnt on CubeSat general system design philosophy, data bus topologies, and management philosophies are discussed and compared against the more traditional small sat approach, something on which the Surrey community can speak with authority. Conclusions are drawn on the the similarities and differences of the small-satellite approach pioneered in the 1990s and the CubeSat approach pioneered in the 2000s, with recommendations on where commercial, small satellite engineering philosophy can be applied to the hypothetical operational CubeSat mission, and vice versa.

Satellite conjunctions in space are highly challenging because of the lack of space situational awareness solutions and orbit data sharing schemes. Two Line Elements set (TLEs) are commonly used to define satellite state on orbit but are highly inaccurate. Similarly, Global Positioning System (GPS) used for positioning and tracking purposes is not, as a standalone solution, optimised for satellite traffic management. Therefore, an autonomous system specifically designed for space traffic management is needed. A new approach has been adapted for different satellite conjunction scenarios and investigated in a way that each satellite is equipped with a radio measurement instrument operating in multiple low-noise bands taking advantage of Software Defined Radio (SDR) concepts. Relative range between satellites has been obtained from the Received Signal Strength (RSS) by implementing adaptive changes in operating frequencies. Doppler frequency shifts have also been obtained which also have a significant importance on tracking satellites. Our results show that for a two-satellite scenario, it is possible to receive a signal with 20 dB signal to noise ratio from approximately 1800 km when operating at High Frequency (HF) and 600 km at Very High Frequency (VHF). Consequently, in a 600-satellite scenario, more than 79 satellites were detected by a main satellite observer when operating at 30 MHz whereas only 10 and 5 satellites were detected when operating at 140.8 MHz and 440.01 MHz respectively. Operating at the higher frequencies (2499 and 5088 MHz) yielded two dangerous close approaches with a maximum relative range of 2 km using both RSS and Doppler. Further, a second method of estimating range based on time of flight (ToF) has been implemented showing directly dependant ranging errors from signal processing and propagation time delays. Combining different ranging methods and altering between transmitting frequencies by using enabling SDR technologies helps to develop a new highly accurate collision detection system which can complement the existing systems and define the nature of satellite conjunctions in space.

The Surrey Training Research and Nanosatellite Demonstrator (STRaND) programme has been success in identifying and creating a leading low-cost nanosatellite programme with advanced attitude and orbit control system (AOCS) and experimental computing platforms based on smart-phone technologies. The next demonstration capabilities, that provide a challenging mission to the existing STRaND platform, is to perform visual inspection, proximity operations and nanosatellite docking. Visual inspection is to be performed using a COTS LIDAR system to estimate range and pose under 100 m. Proximity operations are controlled using a comprehensive guidance, navigation and control (GNC) loop in a polar form of the Hills Clohessy Wiltshire (HCW) frame including J2 perturbations. And finally, nanosatellite docking is performed at under 30 cm using a series of tuned magnetic coils. This paper will document the initial experiments and calculations used to qualify LIDAR components, size the mission thrust and tank requirements, and air cushion table demonstrations of the docking mechanism.

Observations of highly red-shifted 21-cm hydrogen signals have been suggested as the only means to probe the early Universe from recombination to reionization. During this era, called the Dark Ages, the Universe consisted of neutral hydrogen gas and was opaque to light. It did not become transparent, as we see it today, until reionization was completed. The Dark Ages was the time period when matter clumped together, the very first stars and black holes were born, and, eventually, the first galaxies were formed. To enable observations of the Dark Ages is therefore one of the top priorities in cosmology and astrophysics. Today, the cosmological 21-cm signals are highly red-shifted and should peak in the FM radio band. Observing the Dark Ages from Earth is therefore next to impossible, due to man-made radio frequency interference (RFI) and ionospheric disturbances. To efficiently block the RFI, which would otherwise overwhelm the weak cosmological signal; it has been proposed to use the Moon as a radio shield and either place a satellite equipped with an ultra-sensitive radio instrument in lunar orbit or to deploy a large low-frequency radio array on the far-side of the Moon. Such missions are technically challenging and expensive and have so far failed to gain support from any national or international space program. Our goal is therefore to use a constellation of small inexpensive satellites in lunar orbit to collect pathfinder data, which would demonstrate EPSC Abstracts Vol. 9, EPSC2014-798, 2014 European Planetary Science Congress 2014 c Author(s) 2014 EPSC European Planetary Science Congress the feasibility of using the Moon as a radio shield, and map out the spatial extent of this RF quiescent zone to support future missions to explore the cosmos. This paper examines the design and radio payload of this mission. Alternative orbits, constellation and payload designs are analyzed to optimize the mission for performance and cost.

Non-cooperative tracking of shipping from space has traditionally been the preserve of state actors. Recent progress in nanosat and SDR technology raises the possibility of conducting such surveillance from low-cost space platforms. While ships may switch off beacons in order to obscure their location, they cannot proceed safely without their navigation radar systems being operational. Radar emissions may be recorded to achieve geolocation. As the radar antenna is mechanically rotated, a periodic power peak is observed by a fixed receiver. A receiver in LEO will record differing interception times according to its motion and that of the emitter platform. A least squares estimator is used to compare possible emitter locations with the collected data in order to arrive at the most probable geolocation. The highly constrained geometry of the system reduces the search to a 1-dimensional one which significantly reduces the processing power required for a fast answer. It is suggested that power reduction resulting from a new algorithm, coupled with advances in nanosat and SDR technology enables a low-cost demonstrator mission. Additionally, using only one satellite removes the requirement for high-speed data links, further reducing weight and power needs. Accuracies of better than 20km have been recorded in simulations.

This paper is concerned with application of standard wireless COTS protocols to space. Suitability of commercially available wireless sensor mote kits for communication inside and between satellites is investigated Spacecraft applications of motes are being considered and a set of requirements are identified Selected mote kits are tested under various scenarios complying with spacecraft testing procedures. The paper details the results of the carried out functional, EMC/I, vibration, thermal and radiation tests.

Clusters, constellations, formations, or `swarms' of small satellites are fast becoming a way to perform scientific and technological missions more affordably. As objectives of these missions become more ambitious, there are still problems in increasing the number of communication windows, supporting multiple signals, and increasing data rates over reliable intersatellite and ground links to Earth. Also, there is a shortage of available frequencies in the 2 m and 70 cm bands due to rapid increase in the number of CubeSats orbiting the Earth - leading to further regulatory issues. Existing communication systems and radio signal processing Intellectual Property (IP) cores cannot fully address these challenges. One of the possible strategies to solve these issues is by equipping satellites with a Software Defined Radio (SDR). SDR is a key area to realise various software implementations which enable an adaptive and reconfigurable communication system without changing any hardware device or feature. This paper proposes a new SDR architecture which utilises a combination of Field Programmable Gate Array (FPGA) and field programmable Radio Frequency (RF) transceiver to solve back-end and front- end challenges and thereby enabling reception of multiple signals or satellites using single user equipment.

Proximity flight systems for rendezvous-and-docking, are traditionally the domain of large, costly institutional manned missions, which require extremely robust and expensive Guidance Navigation and Control (GNC) solutions. By developing a low-cost and safety compliant GNC architecture and design methodology, low cost GNC solutions needed for future missions with proximity flight phases will have reduced development risk, and more rapid development schedules. This will enable a plethora of on-orbit services to be realised using low cost satellite technologies, and lower the cost of the services to a point where they can be offered to commercial as well as institutional entities and thereby dramatically grow the market for on-orbit construction, in-orbit servicing and active debris removal. It will enable organisations such as SSTL to compete in an area previously exclusive to large institutional players. The AAReST mission (to be launched in 2018), will demonstrate some key aspects of low cost close proximity “co-operative” rendezvous and docking (along with reconfiguration/control of multiple mirror elements) for future modular telescopes. However this is only a very small scale academic mission demonstration using cubesat technology, and is limited to very close range demonstrations. This UK National Space Technology Programme (NSTP-2) project, which is being carried out by SSTL and SSC, is due to be completed by the end of November 2017 and is co-funded by the UK Space Agency and company R&D. It is aiming to build on the AAReST ("Autonomous Assembly of a Reconfigurable Space Telescope") mission (where appropriate), and industrialise existing research, which will culminate in a representative model that can be used to develop low-cost GNC solutions for many different mission applications that involve proximity activities, such as formation flying, and rendezvous and docking. The main objectives and scope of this project are the following:  Definition of a reference mission design (based on a scenario that SSTL considers credible as a realistic scenario) and mission/system GNC requirements.  Develop a GNC architectural design for low cost missions applications that involve close proximity formation flying, rendezvous and docking (RDV&D) - i.e. “proximity activities”  Develop a low cost sensor suite suitable for use on proximity missions  Consider possible regulatory constraints that may apply to the mission The SSTL/SSC reference mission concept is a “co-operative” two-spacecraft rendezvous and docking mission demonstrator using microsatellites (an active Chaser and a passive Target), however the GNC model is generic and can be utilized for other “non-co-operative” rendezvous and docking missions. This paper presents the latest results from the study, particularly the mission analysis, GNC simulation and modelling, sensors, and key mission and spacecraft systems aspects. The results so far show that such a GNC model and mission demonstrator is feasible, and in line with anticipated UK regulatory constraints that may apply to the mission.

This paper is concerned with a satellite sensor network, which applies the concept of terrestrial wireless sensor networks to space. Constellation design and enabling technologies for picosatellite constellations such as distributed computing and intersatellite communication are discussed. The research, carried out at the Surrey Space Centre, is aimed at space weather missions in low Earth orbit (LEO). Distributed satellite system scenarios based on the flower constellation set are introduced. Communication issues of a space based wireless sensor network (SB-WSN) in reference to the Open Systems Interconnection (OSI) networking scheme are discussed. A system-on-a-chip computing platform and agent middleware for SB-WSNs are presented. The system-on-a-chip architecture centred around the LEON3 soft processor core is aimed at efficient hardware support of collaborative processing in SB-WSNs, providing a number of intellectual property cores such as a hardware accelerated Wi-Fi MAC and transceiver core and a Java co-processor. A new configurable intersatellite communications module for picosatellites is outlined.

Abstract Thermospheric density is one of the main sources of uncertainty in the estimation of satellites' position and velocity in low‐Earth orbit. This has negative consequences in several space domains, including space traffic management, collision avoidance, re‐entry predictions, orbital lifetime analysis, and space object cataloging. In this paper, we investigate the prediction accuracy of empirical density models (e.g., NRLMSISE‐00 and JB‐08) against black‐box machine learning (ML) models trained on precise orbit determination‐derived thermospheric density data (from CHAMP, GOCE, GRACE, SWARM‐A/B satellites). We show that by using the same inputs, the ML models we designed are capable of consistently improving the predictions with respect to state‐of‐the‐art empirical models by reducing the mean absolute percentage error (MAPE) in the thermospheric density estimation from the range of 40%–60% to approximately 20%. As a result of this work, we introduce Karman: an open‐source Python software package developed during this study. Karman provides functionalities to ingest and preprocess thermospheric density, solar irradiance, and geomagnetic input data for ML readiness. Additionally, it facilitates developing and training ML models on the aforementioned data and benchmarking their performance at different altitudes, geographic locations, times, and solar activity conditions. Through this contribution, we offer the scientific community a comprehensive tool for comparing and enhancing thermospheric density models using ML techniques. Plain Language Summary Accurately modeling the density of the thermosphere is pivotal for spacecraft operations such as collision avoidance, re‐entry prediction, and orbital lifetime analysis. In this study, our aim is twofold. First, we want to study and compare the performance of data‐driven machine learning (ML) models in predicting thermospheric density data against standard empirical models used in the field, which are used as baseline. By training ML models using precise orbit determination‐derived satellite data, we show that they can achieve significant performance improvement compared to empirical models, with a reduction of 61% in the mean absolute percentage error. Second, we also provide the community with a shared software framework that supports the ingestion of solar irradiance, geomagnetic, and thermospheric density data, as well as a training and benchmarking framework to develop ML models. This framework allows researchers and operators to both train their ML models and to compare them at different periods of the solar cycle, geomagnetic storm conditions, geographical locations, and times. Key Points Machine learning (ML) models can significantly outperform existing physics‐based thermospheric neutral density models on exactly the same inputs ML models can improve over NRLMSISE‐00 and JB‐08 empirical density models by 61% and 39% respectively in the mean absolute percentage error The software allows the creation of ML‐ready data for training and benchmarking new models, supporting solar irradiance and geomagnetic data

The InflateSail CubeSat, designed and built at the Surrey Space Centre (SSC) at the University of Surrey, UK, for the Von Karman Institute (VKI), Belgium, is one of the technology demonstrators for the QB50 programme. The 3.2 kilogram InflateSail is “3U” in size and is equipped with a 1 metre long inflatable boom and a 10 square metre deployable drag sail. InflateSail's primary goal is to demonstrate the effectiveness of using a drag sail in Low Earth Orbit (LEO) to dramatically increase the rate at which satellites lose altitude and re-enter the Earth's atmosphere. InflateSail was launched on Friday 23rd June 2017 into a 505km Sun-synchronous orbit. Shortly after the satellite was inserted into its orbit, the satellite booted up and automatically started its successful deployment sequence and quickly started its decent. The spacecraft exhibited varying dynamic modes, capturing in-situ attitude data throughout the mission lifetime. The InflateSail spacecraft re-entered 72 days after launch. This paper describes the spacecraft and payload, and analyses the effect of payload deployment on its orbital trajectory. The boom/drag-sail technology developed by SSC will next be used on the RemoveDebris mission, which will demonstrate the applicability of the system to microsat deorbiting.

Future space telescopes with diameter over 20 m will require new approaches: either high-precision formation flying or in-orbit assembly. We believe the latter holds promise as a potentially lower cost and more practical solution in the near term, provided much of the assembly can be carried out autonomously. To gain experience, and to provide risk reduction, we propose a combined mico/nano-satellite demonstration mission that will focus on the required optical technology (adaptive mirrors, phase-sensitive detectors) and autonomous rendezvous and docking technology (inter-satellite links, relative position sensing, automated docking mechanisms). The mission will involve two "3U" Cubesat-like nanosatellites ("MirrorSats") each carrying an electrically actuated adaptive mirror, and each capable of autonomous un-docking and re-docking with a small central "15U" class micro/nano-satellite core, which houses two fixed mirrors and a boom-deployed focal plane assembly. All three spacecraft will be launched as a single ∼40kg micro-satellite package. The spacecraft busses are based on heritage from Surrey's SNAP-1 and STRaND-1 missions (launched in 2000 and 2013 respectively), whilst the optics, imaging sensors and shape adjusting adaptive mirrors (with their associated adjustment mechanisms) are provided by CalTech/JPL. The spacecraft busses provide precise orbit and attitude control, with inter-satellite links and optical navigation to mediate the docking process. The docking system itself is based on the electromagnetic docking system being developed at the Surrey Space Centre (SSC), together with rendezvous sensing technology developed for STRaND-2. On orbit, the mission profile will firstly establish the imaging capability of the compound spacecraft before undocking, and then autonomously re-docking a single MirrorSat. This will test the docking system, autonomous navigation and system identification technology. If successful, the next stage will see the two MirrorSat spacecraft undock and re-dock to the core spacecraft in a linear formation to represent a large (but sparse) aperture for high resolution imaging. The imaging of stars is the primary objective, but other celestial and terrestrial targets are being considered. Teams at CalTech and SSC are currently working on the mission planning and development of space hardware. The autonomous rendezvous and docking system is currently under test on a 2D air-bearing table at SSC, and the propulsion and precision attitude control system is currently in development. Launch is planned for 2016. This paper details the mission concept, technology involved and progress to date, focussing on the spacecraft buses.

Human spaceflight to/on/from the Moon will benefit from exploitation of various in-situ resources such as water volatile and mineral. Evidence for water ice in Permanently Shadowed Regions (PSRs) on the Moon is both direct and indirect, and derives from multiple past missions including Lunar Prospector, Chandrayaan-1 and LCROSS. Recent lunar CubeSats missions proposed through the Space Launch Systems (SLS) such as Lunar Flashlight, LunaH-Map and Lunar Ice-Cube, will help improve our understanding of the spatial distribution of water ice in those lunar cold traps. However, the spatial resolution of the observations from these SLS missions is on the order of one to many kilometres. In other words, they can miss smaller (sub-km) surficial deposits or near-surface deposits of water ice. Given that future lunar landers or rovers destined for PSRs will likely have limited mobility (but improved landing precision), there is a need to improve the spatial accuracy of maps of water ice in PSRs. The VMMO (Volatiles and Mineralogy Mapping Orbiter) is a semiautonomous, low-cost 12U lunar Cubesat being developed by a multi-national team funded through European Space Agency (ESA) for mapping lunar volatiles and mineralogy at relatively high spatial resolutions. It has a potential launch in 2023 as part of the ESA/SSTL lunar communications pathfinder orbiter mission. This paper presents the work carried out so far on VMMO concept design and development including objectives, profile, operations and spacecraft payload and bus.

Software Defined Radios (SDRs) have emerged as a viable approach for space communications over the last decade by delivering low-cost hardware and flexible software solutions. The flexibility introduced by the SDR concept not only allows the realization of multiple standards on one platform, but also promises to ease the implementation of one communication standard on differing SDR platforms by waveform porting. This technology would facilitate implementing reconfigurable nodes for parallel satellite reception in Mobile/Deployable Ground Segments. The SDR architecture was implemented initially in C/C++ and tested over varied embedded platforms and at different data rates from 1.2 to 19.2 kbps. Profiling using gprof demonstrated the need to move the up and down sampling blocks demanding higher computation to Field Programmable Gate Array (FPGA) logic in order to benefit new architecture optimization and thereby facilitating more than one signal at any given time. The paper includes the implementation of the Digital Down Converter (DDC) block in VHDL and design tradeoffs that yields insight into optimal solutions along with effective evaluation of the new candidate architecture.

The U.K. CubeSat Forum held a one-day workshop meeting at the Harwell Science and Innovation Campus, Harwell, U.K. in May 2014. One objective of the workshop was for the U.K. CubeSat community – represented by the workshop delegates – to discuss the current and future context for gaining approval, i.e. U.K. government issued licence, to launch and operate a U.K.-registered CubeSat or nanosatellite. This discussion arose given the pre-workshop widespread U.K. community view that to do this within a U.K. context involved significantly more effort, resources and costs than in other countries and perceived to be disproportionate to the overall CubeSat philosophy of low-cost, low-resource and rapid implementation of missions. The workshop attendees (~120 delegates) were split into three parallel discussion groups to discuss this point.

The CubeSat standard has inspired a new wave of engineers, researchers, and scientists – all aiming to utilise „commercial off-the-shelf‟ (COTS) subsystems to build nanosatellite systems. For this same reason, students and staff at Surrey have been designing a 3U CubeSat with the intention of providing low-level design, build and test experience for early career engineers, provide on-orbit demonstration of key technologies in attitude and orbit control systems (AOCS), and finally assess smart-phone components for future spacecraft applications. The modern smart-phone is the latest in a consumer driven and start-of-the-art electronics market which the STRaND-1 mission aims to exploit; assessing hardware components and exploring software towards new methods in designing and operating new satellites. The „Surrey Training, Research and Nanosatellite Demonstrator‟ is the first in a line of ambitious missions which aims to strengthen the ties between Surrey Space Centre (SSC) and Surrey Satellite Technology Ltd (SSTL). The project began in April 2010 where STRaND-1 was designed in the team‟s free time but the major build work began in mid-October 2012. Launched on 25 February 2013, the assembly, integration and test (AIT) phase took approximately 4 months to complete. This paper discusses this phase and the typical-CubeSat subsystems and non-PC/104 AOCS and computing payloads in a custom payload bay. In-orbit results of the new nanosatellite power, attitude and orbit control system are presented. With many new capabilities to demonstrate on STRaND-1 when in-orbit, exploiting the latest consumer electronics and software with real training and mission utility was disruptive. We evaluate the ground-based operational processes during commissioning.

Space researchers at the University of Surrey and Surrey Satellite Technology Limited (SSTL) have developed 'STRaND-1', a satellite containing a smartphone payload that will be launched into orbit around the Earth later this year. STRaND-1 (Surrey Training, Research and Nanosatellite Demonstrator) is being developed by the Surrey team to demonstrate the advanced capabilities of a satellite built quickly using advanced commercial off-the-shelf components. The satellite will be launched into orbit around the Earth in 2011. The phone will run on Android's powerful open-source operating system. A powerful computer, built at the Surrey Space Centre, will test the vital statistics of the phone once in space. The computer will check which components of the phone are working normally and will relay images and messages back to Earth via a radio system. Once all the tests are complete, the plan is to switch off the micro computer and the smartphone will be used to operate parts of the satellite. The smartphone avionics suite is only one of the many technological advances packed into this 4kg satellite. To precisely point and manoeuvre, the satellite also incorporates advanced guidance, navigation and control systems. © 2011 IEEE.

The ongoing evolution in constellation/formation of CubeSats along with steadily increasing number of satellites deployed in Lower Earth Orbit (LEO), demands a generic reconfigurable multimode communication platforms. As the number of satellites increase, the existing protocols combined with the trend to build one control station per CubeSat become a bottle neck for existing communication methods to support data volumes from these spacecraft at any given time. This paper explores the Software Defined Radio (SDR) architecture for the purposes of supporting multiple-signals from multiple-satellites, deploying mobile and/or distributed ground station nodes to increase the access time of the spacecraft and enabling a future SDR for Distributed Satellite Systems (DSS). Performance results of differing software transceiver blocks and the decoding success rates are analysed for varied symbol rates over different cores to inform on bottlenecks for Field Programmable Gate Array (FPGA) acceleration. Further, an embedded system architecture is proposed based on these results favouring the ground station which supports the transition from single satellite communication to multi-satellite communications.

Interest is increasing in the use of neural networks and deep-learning for on-board processing tasks in the space industry [1]. However development has lagged behind terrestrial applications for several reasons: space qualified computers have significantly less processing power than their terrestrial equivalents, reliability requirements are more stringent than the majority of applications deep-learning is being used for. The long requirements, design and qualification cycles in much of the space industry slows adoption of recent developments. GPUs are the first hardware choice for implementing neural networks on terrestrial computers, however no radiation hardened equivalent parts are currently available. Field Programmable Gate Array devices are capable of efficiently implementing neural networks and radiation hardened parts are available, however the process to deploy and validate an inference network is non-trivial and robust tools that automate the process are not available. We present an open source tool chain that can automatically deploy a trained inference network from the TensorFlow framework directly to the LEON 3, and an industrial case study of the design process used to train and optimise a deep-learning model for this processor. This does not directly change the three challenges described above however it greatly accelerates prototyping and analysis of neural network solutions, allowing these options to be more easily considered than is currently possible. Future improvements to the tools are identified along with a summary of some of the obstacles to using neural networks and potential solutions to these in the future.

Future space telescopes with diameter over 20 m will require new approaches: either high-precision formation flying or in-orbit assembly. We believe the latter holds promise at a potentially lower cost and more practical solution in the near term, provided much of the assembly can be carried out autonomously. To gain experience, and to provide risk reduction, we propose a combined micro/nano-satellite demonstration mission that will focus on the required optical technology (adaptive mirrors, phase-sensitive detectors) and autonomous rendezvous and docking technology (inter-satellite links, relative position sensing, automated docking mechanisms). The mission will involve two "3U" CubeSat-like nanosatellites ("MirrorSats") each carrying an electrically actuated adaptive mirror, and each capable of autonomous un-docking and re-docking with a small central "15U" class micro/nano-satellite core, which houses two fixed mirrors and a boom-deployed focal plane assembly. All three spacecrafts will be launched as a single ~40 kg micro-satellite package. The spacecraft busses are based on heritage from Surrey's SNAP-1 and STRaND-1 missions (launched in 2000 and 2013 respectively), whilst the optics, imaging sensors and shape adjusting adaptive mirrors (with their associated adjustment mechanisms) are provided by CalTech/JPL. The spacecraft busses provide precise orbit and attitude control, with inter-satellite links and optical navigation to mediate the docking process. The docking system itself is based on the electromagnetic docking system being developed at the Surrey Space Centre (SSC), together with rendezvous sensing technology developed for STRaND-2. On orbit, the mission profile will firstly establish the imaging capability of the compound spacecraft before undocking, and then autonomously re-docking a single MirrorSat. This will test the docking system, autonomous navigation and system identification technology. If successful, the next stage will see the two MirrorSat spacecraft undock and re-dock to the core spacecraft in a linear formation to represent a large (but sparse) aperture for high resolution imaging. The imaging of stars is the primary objective, but other celestial and terrestrial targets are being considered. Teams at CalTech and SSC are currently working on the mission planning and development of space hardware. The autonomous rendezvous and docking system is currently under test on a 2D air-bearing table at SSC, and the propulsion and precision attitude control system is currently in development. Launch is planned for 2016. This paper details the mission concept; technology involved and progress to date, focussing on the spacecraft buses.

In recent years, there has been a desire to develop space-based optical telescopes with large primary apertures. Current monolithic large telescopes, as exemplified by 6.5m aperture James Webb Space Telescope, are limited by the diameter of the launch vehicle – despite their ability to unfold and deploy mirror elements. One method to overcome this obstacle is to autonomously assemble small independent spacecraft, each with their own mirror, while in orbit. In doing so, a telescope with a large, segmented primary mirror can be constructed. Furthermore, if each of these mirrors is manufactured to have an identical initial shape and then adjusted upon assembly, a substantial reduction in manufacturing costs can be realized. In order to prove the feasibility of such a concept, a collaborative effort between the California Institute of Technology, the University of Surrey, and the Indian Institute of Space Science and Technology has been formed to produce and fly the "Autonomous Assembly of a Reconfigurable Space Telescope" (AAReST) mission. AAReST comprises two 3U Cubesat-like nanosatellites (“MirrorSats”) each carrying an electrically actuated adaptive mirror, and each capable of autonomous un-docking and re-docking with a central “9U” class nanosatellite (“CoreSat”), which houses two fixed mirrors and a boom-deployed focal plane assembly (camera). All three spacecraft will be launched as a single ~30kg microsatellite package. The central premise is that the satellite components can manoeuvre and dock in different configurations and the mirrors can change shape and move to form focused images on the camera focal plane. The autonomous manoeuvres and docking will be under the control of the Surrey developed electro-magnetic docking system and near infra-red lidar/machine-vision based relative navigation sensors. On orbit, the mission profile will firstly establish the imaging capability of the compound spacecraft before undocking, and then autonomously re-docking a single MirrorSat. This will test the docking system, autonomous navigation and system identification technology. If successful, the next stage will see the second MirrorSat spacecraft undock and re-dock to the core spacecraft to form a wide linear formation which represents a large (but sparse) aperture for high resolution imaging. Celestial targets will be imaged. Currently, the flight hardware is under construction and launch is planned for ~2019-2020. This paper details the mission concept, technology involved and its testing and progress on the production of the flight hardware.

—Recently there has been an increasing demand for more powerful processors for the next-generation space missions, such as communication and earth observation. The challenge is how to improve the reliability of the processor under the “single event effects” in orbit. We have previously proposed a new way of implementing any traditional software error detection and correction techniques at instruction level, capable of covering both the CPU and caches of “commercial off the shelf” processors. In this paper, a novel way of evaluation of the software protection is presented, based on a theoretical model and software injection experiments to predict the reliability of the whole processing architecture. The fault injection will evaluate the ability of the protection code to detect and recover errors in addition to the accuracy of the reliability models, by comparing the reliability of the theoretical predictions to the reliability of the injection experiments. Automatic compiler error detection and recovery improves the reliability of the system by reducing the error rate of “single event upsets.” In some benchmarks, the error rate was reduced to less than 1%. This research has been tested in two machines; Intel core i5-3470 and a Raspberry Pi 3. On the first processor, the overhead was less than 15%, and on the second one, the overhead was less than 17%. This research can also be ported to multiple high level languages, with the ability to cover multiple instructions and datatypes

In recent years the growth in quantity, diversity and capability of Earth Observation (EO) satellites, has enabled increase’s in the achievable payload data dimensionality and volume. However, the lack of equivalent advancement in downlink technology has resulted in the development of an onboard data bottleneck. This bottleneck must be alleviated in order for EO satellites to continue to efficiently provide high quality and increasing quantities of payload data. This research explores the selection and implementation of state-of-the-art multidimensional image compression algorithms and proposes a new onboard data processing architecture, to help alleviate the bottleneck and increase the data throughput of the platform. The proposed new system is based upon a backplane architecture to provide scalability with different satellite platform sizes and varying mission’s objectives. The heterogeneous nature of the architecture allows benefits of both Field Programmable Gate Array (FPGA) and Graphical Processing Unit (GPU) hardware to be leveraged for maximised data processing throughput.

The InflateSail (QB50-UK06) CubeSat, designed and built at the Surrey Space Centre (SSC) at the University of Surrey, UK, for the Von Karman Institute (VKI), Belgium – was one of the technology demonstrators for the QB50 pro-gramme. The 3.2 kilogram 3U CubeSat was equipped with a 1 metre long inflat-able boom and a 10m2 deployable drag sail. InflateSail's primary mission was to demonstrate the effectiveness of using a drag sail in Low Earth Orbit (LEO) to dramatically increase the rate at which satellites lose altitude and re-enter the Earth's atmosphere and it was one of 31 satellites that were launched simultane-ously on the PSLV (polar satellite launch vehicle) C-38 from Sriharikota, India on 23rd June 2017 into a 505km, 97.44o Sun-synchronous orbit (SSO). Shortly after orbital insertion, InflateSail booted-up, and, once safely clear of the other satellites on the launch, it automatically activated its payload – firstly, deploying a 1 metre long inflatable boom comprising a metal-polymer laminate tube, using a cool gas generator (CGG) to provide the inflation gas, and secondly, using a brushless DC motor at the end of the boom to extend four lightweight bistable rigid composite (BRC) booms to draw out the 3.1m x 3.1m square, 12 micron thick polymer drag-sail. As intended, the satellite immediately began to lose alti-tude, and re-entered the atmosphere just 72 days later – thus demonstrating for the first time the de-orbiting of a spacecraft using European inflatable and drag-sail technologies. The boom/drag-sail technology developed by SSC will next be used on the RemoveDebris mission, due for launch in 2018, which will demon-strate the capturing and de-orbiting of artificial space debris targets using a net and harpoon system.

Satellite conjunctions in space are a major problem for operators and governments due to the lack of coherent space situational awareness solutions. The tracking accuracy for two-line elements (TLEs) averages in kilometres with similar error boundaries making it limited for critical satellite collision prediction. The common practice using GPS provides high accuracy from centimetres to metres. However, satellite state data (position and velocity) are often never shared and orbit determination methods provide limited solutions at quantifying near-miss events. In the advent of mega-constellations, there is an urgent need for in-situ measurements to develop real satellite traffic management solutions and associated satellite traffic data standardisation to complement and refine the existing techniques. This research presents ToF range estimation techniques adapted for the increasing low Earth orbit satellite traffic that requires co-operative monitoring. Two techniques are investigated namely, two-way time transfer (TWTT) and two-way ranging using direct sequence spread spectrum (TWR-DSSS). Although both techniques reached centimetre-level accuracies (7 to 15 cm) in perfect communications conditions, this accuracy drops quickly when considering the real-world limitations. TWTT technique is affected by processing delay and relative clock drifts. Consequently, the ranging errors standard deviation for TWTT is 210 and 2075 m respectively for the delays 1 and 10 μs. It is also found that the relative clock drifts used for both satellites cause bias ranging errors as the best achieved accuracy is 170 m even when the delays are nullified. On the other hand, TWR-DSSS shows a robust performance against low signal-to-noise (SNR) levels. For instance, relative range is resolved with sub-kilometre accuracy for -20 dB SNR. Ultimately, inter-satellite cooperative RF ranging based on time of flight can offer real opportunities of a new measurement instrument complementing the existing satellite conjunction assessment tools.

CHAFF is a Surrey-designed and built hyperspectral imager prototype intended for the 6U CubeSat platform. Capable of taking hyperspectral images across the wavelength range 460 nm – 820 nm at a best spectral resolution of 3.46 nm (at 546 nm), CHAFF has been designed holistically: consideration of the operational constraints of the CubeSat platform since design inception has allowed the development of techniques which address these constraints within the optical design. For example, CHAFF will employ optically aided image co-registration in order to deal with the physical pointing instability of the CubeSat and co-register the band images on-board the satellite. This will in turn improve the performance of image compression algorithms, which rely on spatial, spectral and statistical redundancy within the hypercube to achieve optimal performance. Additionally, CHAFF has been constructed with commercial off-the-shelf optics, to keep the design commensurate with a university CubeSat budget. Presented here is an update of the progress achieved, focusing on data collected by CHAFF during a field trial undertaken to assess the abilities of the instrument. Performance of the image co-registration and the instrument calibration on natural scene data is assessed, and the behavior of the instrument when presented with vegetation targets is analysed.

While small, low-cost satellites continue to increase in capability and popularity, their reliability remains a problem. Traditional techniques for increasing system reliability are well known to satellite developers, however, their implementation on low-cost satellites is often limited due to intrinsic mass, volume and budgetary restrictions. Aiming for graceful degeneration, therefore, may be a more promising route. To this end, a stem-cell-inspired, multicellular architecture is being developed using commercial-off-the-shelf components. It aims to replace a significant portion of a typical satellite’s bus avionics with a set of initially identical cells. Analogous to biological cells, the artificial cells are able to differentiate during runtime to take on a variety of tasks thanks to a set of artificial proteins. Each cell reconfigures its own proteins within the context of a system-wide distributed task management strategy. In this way, essential tasks can be maintained, even as system cells fail. This paper focusses on two hardware implementations of the stem-cell inspired architecture. The first implementation, based on a single cell, serves as the Payload Interface Computer on a CubeSat named SME-SAT. The second hardware implementation is a benchtop system composed of several cells intended to demonstrate a complete multicellular system in operation. In order to demonstrate the feasibility of these multicellular architectures, the physical attributes of the hardware implementations are compared to those of more traditional implementations and are shown to have enhanced reliability at the cost of increased power and internal bus bandwidth.

Satellite constellation deployment for formation flying missions is one of the key areas for consideration when realizing the final constellation with reduced propellant mass requirements on the propulsion system. The use of a single launch vehicle to deploy multiple satellites into a formation is faster and cheaper but there is greater risk of collision. This risk must be managed with the competing desire to establish a relatively tight formation for better inter-satellite communication. The launcher attitude, satellite injection times and velocities are key parameters to safely achieve a given separation distance and distribution. This paper presents a visual simulator to propagate the satellite trajectories from the launcher using an expanded definition of Hill's equations, and extending to polar relative motion. It is assumed that a simple launcher is used which is incapable of reposition once in orbit. Low injection velocities are exploited to inject large numbers satellites into a stable constellation. Utilizing small tight natural motion formations help to reduce perturbations and the propellant mass required for formation maintenance. SatLauncher is a new visualization tool for investigating the relative motion and key parameters between satellites in these new missions and applications for multi-satellite launchers without the need for any further industrial tool. The QB50 mission is taken forward as a representative scenario requiring our latest software tool and new methods are presented towards collision free formation deployment.

Flight and ground segment software in university missions is often developed only after hardware has matured sufficiently towards flight configuration and also as bespoke codebases to address key subsystems in power, communications, attitude, and payload control with little commonality. This bespoke software process is often hardware specific, highly sequential, and costly in staff/monitory resources and, ultimately, development time. Within Surrey Space Centre (SSC), there are a number of satellite missions under development with similar delivery timelines that have overlapping requirements for the common tasks and additional payload handling. To address the needs of multiple missions with limited staff resources in a given delivery schedule, computing commonality for both flight and ground segment software is exploited by implementing a common set of flight tasks (or modules) which can be automatically generated into ground segment databases to deliver advanced debugging support during system end-to-end test (SEET) and operations. This paper focuses on the development, implementation, and testing of SSC’s common software framework on the Stellenbosch ADCS stack and OBC emulators for numerous missions including Alsat-1N, RemoveDebris, SME-SAT, and InflateSail. The framework uses a combination of open-source embedded and enterprise tools such as the FreeRTOS operating system coupled with rapid development templates used to auto-generate C and Python scripts offline from ‘message databases’. In the flight software, a ‘core’ packet router thread forwards messages between threads for inter process communication (IPC). On the ground, this is complemented with an auto-generated PostgreSQL database and web interface to test, log, and display results in the SSC satellite operations centre. Profiling is performed using FreeRTOS primitives to manage module behaviour, context, time and memory – especially important during integration. This new framework has allowed for flight and ground software to be developed in parallel across SSC’s current and future missions more efficiently, with fewer propagated errors, and increased consistency between the flight software, ground station and project documentation.

Over the next two decades, unprecedented astronomy missions could be enabled by space telescopes larger than the James Webb Space Telescope. Commercially, large aperture space-based imaging systems will enable a new generation of Earth Observation missions for both science and surveillance programs. However, launching and operating such large telescopes in the extreme space environment poses practical challenges. One of the key design challenges is that very large mirrors (i.e. apertures larger than 3m) cannot be monolithically manufactured and, instead, a segmented design must be utilized to achieve primary mirror sizes of up to 100m. Even if such large primary mirrors could be made, it is impossible to stow them in the fairings of current and planned launch vehicles, e.g., SpaceX’s Starship reportedly has a 9m fairing diameter. Though deployment of a segmented telescope via a folded-wing design (as done with the James Webb Space Telescope) is one approach to overcoming this volumetric challenge, it is considered unfeasible for large apertures such as the 25m telescope considered in this study. Parallel studies conducted by NASA indicate that robotic on-orbit assembly (OOA) of these observatories offers the possibility, surprisingly, of reduced cost and risk for smaller telescopes rather than deploying them from single launch vehicles but this is not proven. Thus, OOA of large aperture astronomical and Earth Observation telescopes is of particular interest to various space agencies and commercial entities. In a new partnership with Surrey Satellite Technology Limited and Airbus Defence and Space, the Surrey Space Centre is developing the capability for autonomous robotic OOA of large aperture segmented telescopes. This paper presents the concept of operation and mission analysis for OOA of a 25m aperture telescope operating in the visible waveband of the electromagnetic spectrum; telescopes of this size will be of much value as it would permit 1m spatial resolution of a location on Earth from geostationary orbit. Further, the conceptual evaluation of robotically assembling 2m and 5m telescopes will be addressed; these missions are envisaged as essential technology demonstration precursors to the 25m imaging system.

The size of any single spacecraft is ultimately limited by the volume and mass constraints of currently available launchers, even if elaborate deployment techniques are employed. Costs of a single large spacecraft may also be unfeasible for some applications such as space telescopes, due to the increasing cost and complexity of very large monolithic components such as polished mirrors. The capability to assemble in-orbit will be required to address missions with large infrastructures or large instruments/apertures for the purposes of increased resolution or sensitivity. This can be achieved by launching multiple smaller spacecraft elements with innovative technologies to assemble (or self-assemble) once in space and build a larger much fractionated spacecraft than the individual modules launched. Up until now, in-orbit assembly has been restricted to the domain of very large and expensive missions such as space stations. However, we are now entering into a new and exciting era of space exploitation, where new mission applications/markets are on the horizon which will require the ability to assemble large spacecraft in orbit. These missions will need to be commercially viable and use both innovative technologies and small/micro satellite approaches, in order to be commercially successful, whilst still being safety compliant. This will enable organisations such as SSTL, to compete in an area previously exclusive to large commercial players. However, inorbit assembly brings its own challenges in terms of guidance, navigation and control, robotics, sensors, docking mechanisms, system control, data handling, optical alignment and stability, lighting, as well as many other elements including non-technical issues such as regulatory and safety constraints. Nevertheless, small satellites can also be used to demonstrate and de-risk these technologies. In line with these future mission trends and challenges, and to prepare for future commercial mission demands, SSTL has recently been making strides towards developing its overall capability in “in-orbit assembly in space” using small satellites and low-cost commercial approaches. This includes studies and collaborations with Surrey Space Centre (SSC) to investigate the three main potential approaches for in-orbit assembly, i.e. deployable structures, robotic assembly and modular rendezvous and docking. Furthermore, SSTL is currently developing an innovative small ~20kg nanosatellite (the “Target”) as part of the ELSA-d mission which will include various rendezvous and docking demonstrations. This paper provides an overview and latest results/status of all these exciting recent in-orbit assembly related activities.

Breakthroughs in deep convolutional neural networks for new vision applications in image classification and object detection have pushed forward precision and speed performance indicators in both domains. The future of space exploration relies on the development of novel systems for autonomous operations and onboard data handling especially for computer vision and deep learning. However, previous works on object detection and image classification always operate on the rigid assumption that representative data is available and reliable while merely focusing on offline optimization of architectures for accuracy. This assumption cannot be extended to onboard processing especially in a space environment where unknown scene changes in the visual environment directly affect the performance of machine vision systems. The performance of a deep neural network is as dependent on the input data as it is on the network its self. We propose using a multi-sensory computer vision system that accounts for data reliability and availability using an adaptive input policy. We use custom datasets containing RGB and Depth images of a reference satellite mission for training and testing deep convolutional neural network models for object detection. Our simulation testbed generates our datasets which cover all poses, different ranges, lighting conditions and visual environments. The trained models use multi-sensory input data from both an optical sensor (RGB data) and a time of flight sensor (Depth data). The multi-sensory input data is passed through the adaptive input layer to complementarily provide the most reliable output in a harsh space environment that does not tolerate missing and unreliable data. For instance, the ToF sensor provides visual data that reliably cover close ranges and most importantly can operate regardless of ambient light. The optical sensor provides RGB data at farther ranges and, unlike ToF sensors, is not susceptible to saturation from Earth infra-red emissions. This selective multi-sensory input approach ensures that the CNN model receives reliable input data regardless of the changes in the visual environment to fit the strict operational requirements of space missions. Our work is validated using a sensory-data reliability assessment and object detection models based on the state of the art using Faster R-CNN and YOLO detection techniques. Average precision on the validation dataset saw a significant improvement using our approach. Average precision results went from 50% and 40% using RGB and Depth respectively to 080% using the input selective system.

In this work we propose an adaptive and scalable hardware implementation of convolutional Neural Networks. The adaptive hardware model is the result of a design loop that starts with a software implementation relying on standard scanning window and MAC operations. This design is developed into a deterministic, hardware-friendly model which introduces timing, fixed-point representation and a pixel streaming interface. Then finally HDL code is generated and an RTL of the system is created. Each step is analyzed and validated against pre-set objectives using a golden reference from the last step. The proposed system is capable of selective output execution of different data-paths. It allows for real time trade-offs between accuracy for execution time and power. This is achieved by implementing a CNN network through a number of sequential layer blocks. Layer-blocks can effectively be considered standalone networks with differing complexities. Each layer blocks branches off into an output that is independent of the block that follows it. This allows the system to execute partially or fully according to performance requirements. This reconfigurable model trades off accuracy for speed and power, results show a tradeoff in accuracy for a 50% and 70% gain in both speed and power respectfully.

Training of convolutional neural networks (CNNs) on embedded platforms to support on-device learning has become essential for the future deployment of CNNs on autonomous systems. In this work, we present an automated CNN training pipeline compilation tool for Xilinx FPGAs. We automatically generate multiple hardware designs from high-level CNN descriptions using a multi-objective optimization algorithm that explores the design space by exploiting CNN parallelism. These designs that trade-off resources for throughput allow users to tailor implementations to their hardware and applications. The training pipeline is generated based on the backpropagation (BP) equations of convolution which highlight an overlap in computation. We translate the overlap into hardware by reusing most of the forward pass (FP) pipeline reducing the resources overhead. The implementation uses a streaming interface that lends itself well to data streams and live feeds instead of static data reads from memory. Meaning, we do not use the standard array of processing elements (PEs) approach, which is efficient for offline inference, instead we translate the architecture into a pipeline where data is streamed through allowing for new samples to be read as they become available. We validate the results using the Zynq-7100 on three datasets and varying size architectures against CPU and GPU implementations. GPUs consistently outperform FPGAs in training times in batch processing scenarios, but in data stream scenarios, FPGA designs achieve a significant speedup compared to GPU and CPU when enough resources are dedicated to the learning task. A 2.8x, 5.8x, and 3x speed up over GPU was achieved on three architectures trained on MNIST, SVHN, and CIFAR-10 respectively.

Distributed satellite systems are large research topics, spanning many fields such as communications, networking schemes, high performance computing, and distributed operations. DARPA's F6 fractionated spacecraft mission is a prime example, culminating in the launch of technology demonstration satellites for autonomous and rapidly configurable satellite architectures. Recent developments at Surrey Space Centre have included the development of a Java enabled system-on-a-chip solution towards running homogenous agents and middleware software configurations.

Presented here is CHAFF (CubeSat Hyperspectral Application For Farming), a design concept for a CubeSat-based Hyperspectral Imager (CHSI) intended to supply high quality hyperspectral image data cubes to the agricultural community. CHAFF has been designed holistically as a system, considering all design and operational characteristics of a CHSI instrument and platform together: including the re-stricted payload mass and volume associated with CubeSats, the platform pointing stability/accuracy limitations, and the restricted downlink data budget. To this end, CHAFF will employ optically aided geometric co-registration methods, which will allow on-board construction of the hyperspectral data cube. This allows the use of powerful lossless data compression schemes to mitigate the downlink data budget limitations. In addition, a calibration methodology using a tuneable laser source at NPL, will be employed pre-flight to achieve rapid and accurate spectral and radiometric calibration, essential for the production of science-grade data sets from the proposed CubeSat constellations. A benchtop prototype has been constructed and a promising spectral resolution of 3nm at around 625nm has been achieved. In addition the auxiliary imager for the optically-aided geometric co-registration has been demonstrated.