Current trends in space systems are towards the design of autonomous spacecraft that are capable to generate their own plans of action based on system constraints, on-board data analysis and mission goals set by ground operators. At the same time, the requests in Earth Observation applications has forced the exploration of innovative satellite systems such as satellite swarms, constellations, fully-fractionated spacecraft and federated satellite systems. Enabled by miniaturization techniques and small-spacecraft technologies (e.g. nano-satellites), these distributed mission architectures present improved system qualities (i.e. modularity, scalability, resiliency, incremental deployment...) and performance (e.g. spatial resolution, shorter revisit times...). Although Mission Planning Systems (MPS) for satellites are traditionally ground-based and centralized, providing autonomous capabilities to distributed spacecraft could be the only feasible solution in these new mission architectures where requests are served in a cooperative manner. Regardless of the benefits of autonomous distributed spacecraft, these approaches pose significant challenges in the core component of their planning systems, namely, distributed task schedulers. In this context, this thesis will contribute to the exploration of task schedulers for distributed spacecraft. The purpose of this thesis is twofold: on the one hand to fully design and implement a distributed task scheduling policy. This approach is targeted for the broad range of distributed satellite missions, presenting an adaptive management policy based on the collaboration between two levels of control for dynamic environments with limited computational capabilities. On the other hand the performance of this policy is tested in a simulated environment, against another state-of-the-art alternative, which has been also completely implemented. A benchmark framework has been specifically designed for this purpose. The obtained results have allowed to detect the weaknesses and strengths of this policy and the next steps to be taken in order to have a distributed task scheduler prepared for the future distributed spacecraft architectures.
In the early 1990s, NASA Goddard Space Flight Center started researching and developing autonomous and autonomic ground and spacecraft control systems for future NASA missions. This research started by experimenting with and developing expert systems to automate ground station software and reduce the number of people needed to control a spacecraft. This was followed by research into agent-based technology to develop autonomous ground c- trol and spacecraft. Research into this area has now evolved into using the concepts of autonomic systems to make future space missions self-managing and giving them a high degree of survivability in the harsh environments in which they operate. This book describes much of the results of this research. In addition, it aimstodiscusstheneededsoftwaretomakefutureNASAspacemissionsmore completelyautonomousandautonomic.Thecoreofthesoftwareforthesenew missions has been written for other applications or is being applied gradually in current missions, or is in current development. It is intended that this book should document how NASA missions are becoming more autonomous and autonomic and should point to the way of making future missions highly - tonomous and autonomic. What is not covered is the supporting hardware of these missions or the intricate software that implements orbit and at- tude determination, on-board resource allocation, or planning and scheduling (though we refer to these technologies and give references for the interested reader).
Timely and well-based decision-making plays a crucial role in improving the performance, success rate, and safety of autonomous aerospace systems, especially for those involving multi-agent networks, hybrid dynamical systems, and operations under uncertain environments. Many space missions benefit from the performance and resilience improvements that employing optimal decision-making strategies significantly increases autonomy level, maneuverability, and multi-task capability. However, in real-world applications, a wide range of aerospace systems involve decisions at different autonomy levels and they are usually coupled with each other, which makes it challenging to find the optimal mixed-variable decisions. Optimization and optimal control are the main tools in decision-making. This dissertation aims to develop a systemic rank-constrained optimization methodology for the decision-making of autonomous systems in space missions where the system states are represented by mixed discrete and continuous variables. Rank constrained optimization is to optimize a convex function subject to a convex set of constraints and a rank constraint on the unknown matrix. It has received increasing attention in the areas of matrix completion, signal processing, and model reduction, just to name a few. However, the connection between rank-constrained optimization, especially for rank one-constrained optimization, and mixed-variable decision-making problems has not been well established. In fact, any discrete variable can be regarded as a continuous variable with a polynomial equality constraint. Meanwhile, many system dynamics can be converted into polynomial constraints through discretization and conversion of expressions. Thus, a mixed-variable decision-making problem could be cast as a polynomial optimization problem, which can be expressed as an equivalent quadratically constrained quadratic programming (QCQP) problem by introducing extra variables and quadratic equalities. Furthermore, a general QCQP can be equivalently transformed into a linear matrix programming problem by introducing a to-be-determined rank-one matrix. This dissertation focuses on establishing computationally efficient programs to solve the resulting rank constrained optimization problems and to evaluate the effectivity, efficiency, and performance of the proposed methodologies in decision-making for autonomous systems stemming from applications closely related to space missions. The products contain (1) development of a unified modeling route to formulate a decision-making problem to a general QCQP or rank-constrained optimization problems (RCOP), (2) proposition of four different sequential algorithms, named alternating minimization algorithm (AMA) combined with penalty function, Alternating Projection Approach (APA), Alternating Rank Minimization Approach (ARMA), and Customized Alternating Direction Method of Multipliers (ADMM), to solve the resulting QCQP or RCOP, (3) applications in space missions including a mission planning for spacecraft rendezvous and docking mission, and a fuel optimal guidance for Mars entry, powered descent, and landing mission, (4) applications in other areas including a sensor localization mission of a multi-agent system, and a UAV path-planning problem. Work in this dissertation removes a computational bottleneck in solving a broad class of challenging mixed-variable optimization problems using a uniform formulation associated with a standard routine. The research products, composed of theoretical analysis, algorithm developments, and practical applications, collectively contribute to the full autonomy of a wide class of autonomous systems in space missions.
Advanced space exploration is performed by unmanned missions with integrated autonomy in both flight and ground systems. Risk and feasibility are major factors supporting the use of unmanned craft and the use of automation and robotic technologies where possible. Autonomy in space helps to increase the amount of science data returned from missions, perform new science, and reduce mission costs. Elicitation and expression of autonomy requirements is one of the most significant challenges the autonomous spacecraft engineers need to overcome today. This book discusses the Autonomy Requirements Engineering (ARE) approach, intended to help software engineers properly elicit, express, verify, and validate autonomy requirements. Moreover, a comprehensive state-of-the-art of software engineering for aerospace is presented to outline the problems handled by ARE along with a proof-of-concept case study on the ESA's BepiColombo Mission demonstrating the ARE’s ability to handle autonomy requirements.
Distributed satellite systems is an enabling technology for many future NASA/DoD earth and space science missions, such as MMS, MAXIM, Leonardo, and LISA [1, 2, 3]. While formation flying offers significant science benefits, to reduce the operating costs for these missions it will be essential that these multiple vehicles effectively act as a single spacecraft by performing coordinated observations. Autonomous guidance, navigation, and control as part of a coordinated fleet-autonomy is a key technology that will help accomplish this complex goal. This is no small task, as most current space missions require significant input from the ground for even relatively simple decisions such as thruster burns. Work for the NMP DS1 mission focused on the development of the New Millennium Remote Agent (NMRA) architecture for autonomous spacecraft control systems. NMRA integrates traditional real-time monitoring and control with components for constraint-based planning, robust multi-threaded execution, and model-based diagnosis and reconfiguration. The complexity of using an autonomous approach for space flight software was evident when most of its capabilities were stripped off prior to launch (although more capability was uplinked subsequently, and the resulting demonstration was very successful). How, Jonathan P. and Campbell, Mark and Dennehy, Neil (Technical Monitor) Goddard Space Flight Center MIT-OSP-6891850
High-Reliability Autonomous Management Systems for Spacecraft updates on research on three levels of self-management, including: 1) Autonomous health management of spacecraft that covers how spacecraft can monitor their own state and autonomously detect, isolate and recover from faults; 2) Autonomous mission management of spacecraft where the spacecraft can directly receive the mission, formulate a reasonable plan according to the current state and working environment of the spacecraft, and convert the mission into a specific sequence of instructions; 3) Spacecraft autonomous data management where the spacecraft processes a large amount of raw data and extracts useful information, and autonomously executes or changes flights. The autonomous management of spacecraft uses modern control technologies such as artificial intelligence to establish a remote intelligent body on the spacecraft so that the spacecraft can complete its flight tasks by itself. Its goal is to accurately perceive its own state and external environment without relying on external information injection and control or rely on external control as little as possible. Divides the autonomous management level of spacecraft into two levels, autonomy and execution Covers the implementation of spacecraft autonomous management into three aspects, including autonomous health management of the spacecraft, mission management, and converting the mission into a specific sequence of instructions Discusses how these processes can take a large amount of raw data and extract useful information Covers the autonomous management model of the spacecraft, including compatibility
Autonomous navigation refers to satellites performing on-board, real-time navigation without external input. As satellite systems evolve into more distributed architectures, autonomous navigation can help mitigate challenges in ground operations, such as determining and disseminating orbit solutions. Several autonomous navigation methods have been previously studied, using some combination of on-board sensors that can measure relative range or bearing to known bodies, such as horizon and star sensors (Hicks and Wiesel, 1992) or magnetometers and sun sensors (Psiaki, 1999), however these methods are typically limited to low Earth orbit (LEO) altitudes or other specific orbit cases. Another autonomous navigation method uses intersatellite data, or direct observations of the relative position vector from one satellite to another, to estimate the orbital positions of both spacecraft simultaneously. The seminal study of the intersatellite method assumes the use of radio time-of-flight measurements of intersatellite range, and a visual tracking camera system for measuring the inertial bearing from one satellite to another (Markley, 1984). Due to the limited range constraints of passively illuminated visual tracking systems, many of the previous studies restrict the separation between satellites to less than 1,000 kilometers (e.g., Psiaki, 2011), or simply drop the use of measuring intersatellite bearing and rely solely on obtaining a large distribution of intersatellite range measurements for state estimation (e.g., Xu et al., 2014). These assumptions have limited the assessment of the performance capability of autonomous navigation using intersatellite measurements for more general mission applications. In this thesis, we investigate the performance of using laser communication (lasercom) crosslinks in order to achieve all necessary intersatellite measurements for autonomous navigation. Lasercom systems are capable of measuring both range and bearing to a receiving terminal with greater precision than traditional methods, and can do so over larger separations between satellites. We develop a simulation framework to model the measurements of intersatellite range and bearing using lasercom crosslinks in distributed satellite systems, with consideration of varying orbital operating environments, constellation size and distribution, and network topologies. We implement two estimation algorithms: an extended Kalman filter (EKF) used with Monte Carlo sampling for performance analyses, and a Cram~r-Rao lower-bound (CRLB) computation for uncertainty analyses. We evaluate several case studies modeled off of existing and planned constellation missions in order to demonstrate the new capabilities of performing intersatellite navigation with lasercom links in both near-Earth and deep-space orbital applications. Performance targets are generated from the current state-of-the-art navigation capabilities demonstrated by Global Navigation Satellite Systems (GNSS) in Earth-orbit, and by radiometric tracking and orbit estimation using the Deep Space Network (DSN) in deep-space orbits. For Earth-orbiting applications, we simulate a relay satellite system in geosynchronous orbit (GEO) inspired by current optical communications missions in development by both ESA and NASA, and Walker constellations in LEO inspired by the upcoming mega-constellations for global broadband internet service, such as those proposed by SpaceX and Telesat. In both case studies, we demonstrate improved navigation performance over the current state-of-the-art in GNSS receiver technologies by using intersatellite measurements from lasercom crosslinks. Monte Carlo simulations show median total position errors better than 3 meters in LEO, 12 meters in GEO, and 45 meters in high-altitude or highly-eccentric orbits (HEO), showing promise as an alternative navigation method to GNSS in Earth-orbiting environments. We also simulate current and future Mars-orbiting missions as examples of deep-space applications. In one case study, we create an ad-hoc constellation comprised of low-altitude Mars exploration orbiters modeled off of current Mars-orbiting missions. In a second case study, we focus on a high-altitude constellation proposed for dedicated Earth-to-Mars networked communications. Again, in both case studies, we demonstrate improved navigation performance over the current state-of-the-art in DSN radiometric orbit solutions by using intersatellite measurements from lasercom crosslinks. Monte Carlo simulations show stable median total position errors better than 25 meters for Mars-orbit, which demonstrates a notable improvement both spatially and temporally versus DSN orbit estimation, mitigating the large cost and time constraints associated with DSN tracking. These results demonstrate the promise of using lasercom intersatellite links for autonomous navigation, offering enhanced performance over current state-of-the-art capabilities, and a greater applicability to missions both near Earth and beyond.
This book introduces readers to the fundamentals of estimation and dynamical system theory, and their applications in the field of multi-source information fused autonomous navigation for spacecraft. The content is divided into two parts: theory and application. The theory part (Part I) covers the mathematical background of navigation algorithm design, including parameter and state estimate methods, linear fusion, centralized and distributed fusion, observability analysis, Monte Carlo technology, and linear covariance analysis. In turn, the application part (Part II) focuses on autonomous navigation algorithm design for different phases of deep space missions, which involves multiple sensors, such as inertial measurement units, optical image sensors, and pulsar detectors. By concentrating on the relationships between estimation theory and autonomous navigation systems for spacecraft, the book bridges the gap between theory and practice. A wealth of helpful formulas and various types of estimators are also included to help readers grasp basic estimation concepts and offer them a ready-reference guide.
There is a growing interest in future space mission concepts which involve the interaction of multiple satellites, including on-orbit servicing and debris mitigation, space situational awareness, and swarm-based sensing. These so-called distributed space systems place strict requirements on spaceborne relative navigation accuracy and robustness, autonomy in early phases of the mission, scalability to multiple agents, and resource efficiency in order to meet mission specifications. This research rises to these demands by focusing on a mid- to far-range estimation method called angles-only navigation, wherein observer satellites use bearing angles obtained from onboard monocular camera imagery of target space objects to infer the target orbital trajectories. In this context, vision-based estimation is chosen because it provides a passive and high-dynamic-range navigation technology that uses simple, flight-proven, and miniaturization-friendly hardware. However, since a target appears as only a cluster of pixels in the images, estimating its orbital motion with respect to the observer is fundamentally constrained due to a lack of range information and generally results in poor (or even divergent) navigation performance. To overcome this limitation, a novel angles-only navigation architecture is developed which leverages a deep insight into the relative motion dynamics and advanced filtering techniques to capture key nonlinearities in the dynamics and measurement modeling that lead to range disambiguation. Whereas current approaches require repetition of complex maneuver profiles to gain new vantage points on the target for range rectification, the new methods posed here are completely maneuver-free. To address further deficiencies in the current state of the art, the proposed framework generalizes the applicability of angles-only navigation to new domains beyond low Earth orbit, to situations where no prior state knowledge is available for estimator initialization, and to new scenarios involving multiple observers and/or targets. The functionality and performance of the proposed navigation architecture are verified using rigorous software-based and hardware-in-the-loop simulation. Finally, the algorithms developed in this work have led to the systematic design of the Starling Formation-flying Optical eXperiment (StarFOX), which will test angles-only navigation for future deep-space swarms while flying on board NASA's Starling1 technology demonstration formation in 2021.
Satellite imagery has become an essential resource for environmental, humanitarian, and industrial endeavours. As a means to satisfy the requirements of new applications and user needs, novel Earth Observation (EO) systems are exploring the suitability of Distributed Satellite Systems (DSS) in which multiple observation assets concurrently sense the Earth. Given the temporal and spatial resolution requirements of EO products, DSS are often envisioned as large-scale systems with multiple sensing capabilities operating in a networked manner. Enabled by the consolidation of small satellite platforms and fostered by the emerging capabilities of distributed systems, these new architectures pose multiple design and operational challenges. Two of them are the main pillars of this research, namely, the conception of decision-support tools to assist the architecting process of a DSS, and the design of autonomous operational frameworks based on decentralised, on-board decision-making.The first part of this dissertation addresses the architecting of heterogeneous, networked DSS architectures that hybridise small satellite platforms with traditional EO assets. We present a generic design-oriented optimisation framework based on tradespace exploration methodologies. The goals of this framework are twofold: to select the most optimal constellation design; and to facilitate the identification of design trends, unfeasible regions, and tensions among architectural attributes. Oftentimes in EO DSS, system requirements and stakeholder preferences are not only articulated through functional attributes (i.e. resolution, revisit time, etc.) or monetary constraints, but also through qualitative traits such as flexibility, evolvability, robustness, or resiliency, amongst others. In line with that, the architecting framework defines a single figure of merit that aggregates quantitative attributes and qualitative ones-the so-called ilities of a system. With that, designers can steer the design of DSS both in terms of performance or cost, and in terms of their high-level characteristics. The application of this optimisation framework has been illustrated in two timely use-cases identified in the context of the EU-funded ONION project: a system that measures ocean and ice parameters in Polar regions to facilitate weather forecast and off-shore operations; and a system that provides agricultural variables crucial for global management of water stress, crop state, and draughts.The analysis of architectural features facilitated a comprehensive understanding of the functional and operational characteristics of DSS. With that, this thesis continues to delve into the design of DSS by focusing on one particular functional trait: autonomy. The minimisation of human-operator intervention has been traditionally sought in other space systems and can be especially critical for large-scale, structurally dynamic, heterogeneous DSS. In DSS, autonomy is expected to cope with the likely inability to operate very large-scale systems in a centralised manner, to improve the science return, and to leverage many of their emerging capabilities (e.g. tolerance to failures, adaptability to changing structures and user needs, responsiveness). We propose an autonomous operational framework that provides decentralised decision-making capabilities to DSS by means of local reasoning and individual resource allocation, and satellite-to-satellite interactions. In contrast to previous works, the autonomous decision-making framework is evaluated in this dissertation for generic constellation designs the goal of which is to minimise global revisit times. As part of the characterisation of our solution, we stressed the implications that autonomous operations can have upon satellite platforms with stringent resource constraints (e.g. power, memory, communications capabilities) and evaluated the behaviour of the solution for a large-scale DSS composed of 117 CubeSat-like satellite units.
