Autonomous spacecraft making independent high-level decisions present the promise of dramatically increased productivity in space for both exploration and economic activity. While autonomy has seen limited use in space to date owing to a lack of flight heritage, limited computational resources, and a traditionally risk adverse industry, the growing numbers of spacecraft and increasingly ambitious missions will soon render the current ground-intensive mode of space operation untenable.
\r\n \r\nIn this thesis, we develop two critical capabilities for an autonomous future in space. The first is proactive fault estimation, which seeks to rapidly and safely identify the root causes of onboard anomalies by planning sequences of test actions to gather information while probabilistically ensuring safety. The second is real-time reconfiguration to enable formations of spacecraft to respond quickly and effectively to changing environments or mission objectives.
\r\n\r\nWe achieve both goals using various forms of Monte-Carlo Tree Search planning. By formalizing each capability as sequential decision-making problems, and developing algorithms well suited to information gathering, we show that our algorithms provably converge to optimal solutions while maintaining the ability to run in real-time on robotic spacecraft simulators. We present several algorithmic innovations, including marginalized filtering, sampling-based chance constraint evaluation, and an array-based implementation of Monte-Carlo Tree Search. Through and numerical simulations and hardware experiments, we demonstrate that these modifications enable our algorithms to outperform existing tree search methods and achieve better scaling across system complexity, noise, and simulation depth.
", "doi": "10.7907/ptpk-d504", "publication_date": "2025", "thesis_type": "phd", "thesis_year": "2025" }, { "id": "thesis:16485", "collection": "thesis", "collection_id": "16485", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06032024-181240783", "primary_object_url": { "basename": "connor-lee-thesis.pdf", "content": "final", "filesize": 31343232, "license": "other", "mime_type": "application/pdf", "url": "/16485/1/connor-lee-thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Learning-Based Perception for Robotics in Suboptimal Data Landscapes", "author": [ { "family_name": "Lee", "given_name": "Connor Tinghan", "orcid": "0000-0002-5008-4092", "clpid": "Lee-Connor-Tinghan" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Watkins", "given_name": "Michael M.", "clpid": "Watkins-M-M" }, { "family_name": "Gkioxari", "given_name": "Georgia", "clpid": "Gkioxari-Georgia" }, { "family_name": "Hadaegh", "given_name": "Fred Y.", "orcid": "0000-0002-0992-6323", "clpid": "Hadaegh-F-Y" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Autonomous robots are increasingly present in the world today, being used across a variety of settings and applications. In order to interact with their surroundings, robots typically use cameras to see the world, employing computer vision algorithms to comprehend rich, visual information. While contemporary, learning-based computer vision models provide robots with an accurate and robust understanding of their surroundings, most off-the-shelf methods rely on supervised deep learning techniques, requiring abundant labeled data in order to train and prevent overfitting. However, in many robotic applications and settings, the data landscape is characterized by data scarcity and/or the lack of apparent supervisory signals. Since custom perception solutions are often required for robotic applications, direct adoption of common computer vision methods proves challenging.
\r\n\r\nIn this thesis, we develop robotic perception approaches across three different applications that overcome the challenges of such data landscapes. First, we develop learning-based visual terrain-relative navigation (VTRN) approaches for high-altitude aerial vehicles. This is a problem for which relevant data is available, but made difficult by the lack of obvious supervisory signals related to the high-level navigation objective. In the first chapters of the thesis, we show the power of self-supervised learning approaches to increase VTRN robustness to seasonal and temporal variations that would otherwise debilitate such systems.
\r\n\r\nNext, we address the challenge of developing thermal semantic perception algorithms for aerial field robotics. Due to the specialized nature of field environments and the sensing modality, development of thermal vision algorithms under these conditions is often characterized by the lack of relevant data. We show how we develop various thermal semantic segmentation in response to the evolving data constraints inherent in field robotic projects. In the final part of the thesis, we develop data-efficient, multispectral deep learning algorithms for autonomous driving applications where the lack of data arises from the need for custom, multispectral datasets that are synchronized and coregistered.
", "doi": "10.7907/v4yf-pj25", "publication_date": "2024", "thesis_type": "phd", "thesis_year": "2024" }, { "id": "thesis:16482", "collection": "thesis", "collection_id": "16482", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06032024-152357240", "type": "thesis", "title": "Do Robots Dream of Random Trees? Monte Carlo Tree Search for Dynamical, Partially Observable, and Multi-Agent Systems", "author": [ { "family_name": "Rivi\u00e8re", "given_name": "Benjamin Pierre", "orcid": "0000-0002-0597-5400", "clpid": "Rivi\u00e8re-Benjamin-Pierre" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Yue", "given_name": "Yisong", "orcid": "0000-0001-9127-1989", "clpid": "Yue-Yisong" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" }, { "family_name": "Hadaegh", "given_name": "Fred", "orcid": "0000-0002-0992-6323", "clpid": "Fred-Hadaegh" }, { "family_name": "Pellegrino", "given_name": "Sergio", "orcid": "0000-0001-9373-3278", "clpid": "Pellegrino-S" } ], "local_group": [ { "literal": "GALCIT" }, { "literal": "div_eng" } ], "abstract": "Autonomous robots are poised to transform various aspects of society, spanning transportation, labor, and scientific space exploration. A critical component to enable their capabilities is the algorithm that interprets sensor data to generate intelligent planned behavior. Although reinforcement learning methods that train parameterized policies offline from data have shown recent success, they are inherently limited when robots inevitably encounter situations outside their training domain. In contrast, optimal control techniques, which compute trajectories in real-time using numerical optimization, typically yield only locally optimal solutions.
\r\n\r\nThis research endeavors to bridge the gap by developing algorithms that compute trajectories in real-time while converging towards globally optimal solutions. Building upon the Monte Carlo Tree Search (MCTS) framework\u2014a stochastic tree search method that simulates future trajectories while balancing exploration and exploitation\u2014the research focus is twofold: (i) constructing an efficient discrete representation of continuous systems in a decision trees, and (ii) searching on the resulting tree while balancing exploration and exploitation to achieve global optimality.
\r\n\r\nThe study spans theoretical analysis, algorithmic design, and hardware demonstrations across dynamical, partially observable, and multi-agent systems. By addressing these critical questions, this research aims to advance the field of autonomous robotics, enabling the deployment of intelligent robots in complex and diverse environments.
", "doi": "10.7907/dbwa-we50", "publication_date": "2024", "thesis_type": "phd", "thesis_year": "2024" }, { "id": "thesis:16069", "collection": "thesis", "collection_id": "16069", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06022023-235550987", "primary_object_url": { "basename": "Thesis_Kai_Matsuka.pdf", "content": "final", "filesize": 14316975, "license": "other", "mime_type": "application/pdf", "url": "/16069/1/Thesis_Kai_Matsuka.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Vision-Based Navigation and Large-Scale Estimation for Spacecraft Swarms", "author": [ { "family_name": "Matsuka", "given_name": "Kai", "orcid": "0000-0003-2116-9756", "clpid": "Matsuka-Kai" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Watkins", "given_name": "Michael M.", "clpid": "Watkins-M-M" }, { "family_name": "Hadaegh", "given_name": "Fred", "orcid": "0000-0002-0992-6323", "clpid": "Fred-Hadaegh" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "There has been dramatic growth in the space industry over the past 20 years. Around the same time, robotics and autonomy research has advanced significantly, resulting in a plethora of new mission concepts employing autonomy, such as on-orbit inspection, mission extension, space structure assembly, and orbital debris removal becoming within the realm of possibility. Two of the key autonomous technologies that are critical to the success of these missions are (1) advanced coordination of multi-agent systems and (2) robust vision-based navigation for on-orbit servicing in close proximity. However, there are challenges to simply applying the existing technology to space systems. First, there are domain-specific challenges that are unique to space, such as orbital mechanics and harsh lighting conditions. Second, even at a theoretical level, previous works in the controls and robotics literature are limited when applied to large-scale, locally coupled systems such as spacecraft swarms. To this end, this thesis develops novel algorithms for addressing these gaps.
\r\n\r\nIn the first part of the thesis, we present a decentralized, scalable algorithm for swarm localization, called the Decentralized Pose Estimation (DPE) algorithm. With the DPE algorithm, each spacecraft computes relative navigation estimates with respect to others in the swarm but achieves higher performance through the benefit of multi-agent coordination. The DPE algorithm considers both communication and relative sensing graphs and defines an observable local formation. Each spacecraft jointly localizes its local subset of spacecraft using direct and communicated measurements. Since the algorithm is local, the algorithm complexity does not grow with the number of spacecraft in the swarm. As part of the DPE, we present the Swarm Reference Frame Estimation (SRFE) algorithm, a distributed consensus algorithm to co-estimate a common Local-Vertical, Local-Horizontal frame. The DPE combined with the SRFE provides a scalable, fully-decentralized navigation solution that improves the estimation accuracy compared to when without multi-agent coordination. Numerical simulations and experiments using Caltech's robotic spacecraft simulators are presented to validate the effectiveness and scalability of the DPE algorithm. We show that DPE has much higher accuracy than the best possible estimate without any coordination, while simultaneously being scalable to an arbitrarily large number of agents.
\r\n\r\nIn the second part of the thesis, we propose a novel computer vision algorithm to track the pose of an unknown and uncooperative target using multiple decentralized observers. Vision-based pose determination of an unknown target is challenging due to factors such as lack of cooperative visual markers and harsh lighting conditions of space, and the problem is even harder for distributed observers. To address this challenge, we develop the algorithm called the Multi-Spacecraft Simultaneous Estimation of Pose and Shape algorithm or MSEPS. Within MSEPS, a team of chaser spacecraft, each equipped with a monocular camera, exchange information over a local network to jointly estimate the relative kinematic state of the target and its sparse shape landmarks. In this approach, each spacecraft processes its images and extracts its own set of visual keypoints in parallel. Then, the team uses the local network to jointly estimate the target pose and shape in a distributed fashion by applying the consensus algorithm over the inter-spacecraft communication links. To the best of the authors' knowledge, this is the first cooperative vision-based algorithm for estimating the pose and shape of a space object by means of an arbitrary number of spacecraft.\r\nWe validate our algorithm using simulations of relative orbits and observations captured by each chaser spacecraft and show the multiple observers successfully agree on a consistent estimate and track the target pose accurately.
\r\n\r\nIn the third part of the thesis, we develop some new simulation tools that bridge the gap between robotics and space technology. When developing robotics algorithms for on-orbit systems such as DPE and MSEPS, we identified a need for new simulation tools that tightly integrate robotics algorithms with high-fidelity models of space environments such as astrodynamics effects and visual conditions. To this end, we first develop a ROS2-compatible software interface for Basilisk, the open-source astrodynamics simulation software. This tool allows running Basilisk in parallel with ROS2 in real-time and translates messages between Basilisk modules and ROS2 modules, such that control algorithms implemented in ROS2 can interact with the high-fidelity dynamics within Basilisk in a closed-loop fashion. Second, we develop a ROS2-compatible camera simulation module that uses the Neural Radiance Fields (NeRF) to rapidly generate novel images. These synthetic images are used as inputs to validate the vision-based navigation algorithm in a closed-loop fashion. To validate these simulation tools, we also developed a set of autonomous algorithms for on-orbit inspection and use the simulated measurements as inputs to the algorithm. The real-time numerical simulations demonstrate that our tools can be integrated with autonomy algorithms implemented in ROS2 in a closed-loop fashion to validate the feasibility of the mission.
\r\n\r\nIn the process of addressing some lessons learned from DPE and MSEPS works, we identified that there is a gap in general frameworks for solving the optimal estimation problems for probabilistic inference of large-scale problems involving networked systems. This gap is not just applicable to spacecraft swarms, but also to a general class of large-scale, multi-agent problems in robotics and controls such as localization and mapping, wireless sensor networks, and electrical power grids. Therefore, in the fourth part of the thesis, we address this fundamental gap by developing novel algorithms for Distributed Factor Graph Optimization (DFGO) problems that arise in large-scale networked systems. We develop algorithms for both batch and real-time problems. First, for the batch DFGO problem, we derive a type of the Alternating Direction Method of Multipliers (ADMM) algorithm called the Local Consensus ADMM (LC-ADMM). LC-ADMM is fully localized; therefore, the computational effort, communication bandwidth, and memory for each agent scale like O(1) with respect to the network size. We establish two new theoretical results for LC-ADMM: (1) exponential convergence when the objective is strongly convex and has a Lipschitz continuous subdifferential, and (2) o(1/k) when the objective is convex and has a unique solution. We also show that LC-ADMM allows the use of non-quadratic loss functions, such as l1-norm and Huber loss. Second, we also develop the Incremental DFGO algorithm (iDFGO) for real-time problems by combining the ideas from LC-ADMM and the Bayes tree. To derive a time-scalable algorithm, we exploit the temporal sparsity of the real-time factor graph and the convergence of the augmented factors of LC-ADMM. The iDFGO algorithm incrementally recomputes estimates when new factors are added to the graph and is scalable with respect to both network size and time. We validate LC-ADMM and iDFGO in simulations with examples from multi-agent Simultaneous Localization and Mapping (SLAM) and power grids.
", "doi": "10.7907/spf8-8p84", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:16091", "collection": "thesis", "collection_id": "16091", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06072023-134620248", "primary_object_url": { "basename": "o'connell_michael_2023.pdf", "content": "final", "filesize": 20683543, "license": "other", "mime_type": "application/pdf", "url": "/16091/1/o'connell_michael_2023.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Methods for Robust Learning-Based Control", "author": [ { "family_name": "O'Connell", "given_name": "Michael Thomas", "orcid": "0000-0001-6681-8823", "clpid": "O'Connell-Michael-Thomas" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Yue", "given_name": "Yisong", "orcid": "0000-0001-9127-1989", "clpid": "Yue-Yisong" }, { "family_name": "Burdick", "given_name": "Joel Wakeman", "orcid": "0000-0002-3091-540X", "clpid": "Burdick-J-W" }, { "family_name": "Pellegrino", "given_name": "Sergio", "orcid": "0000-0001-9373-3278", "clpid": "Pellegrino-S" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis addresses the general problem of improving control, safety, and reliability of multi-rotor drones in various challenging conditions by introducing novel deep-learning-based approaches. These approaches are designed to tackle specific issues that multi-rotor drones face during operation, such as near-ground trajectory control, high-speed wind disturbances, actuation delays, and motor failures. The thesis is organized into four main chapters, plus an introduction and conclusion. Each of the main chapters focuses on a unique approach to address a particular challenge of deep-learning-based control methods. Chapter 2 presents Neural-Lander, a deep-learning-based robust nonlinear controller that significantly improves quadrotor control performance during landing by accounting for complex aerodynamic effects. This chapter addresses key challenges to incorporating learned residual dynamics into a control architecture, laying the groundwork for the subsequent chapters. Chapters 3 and 4 introduce Neural-Fly, a learning-based approach that uses Domain Adversarially Invariant Meta-Learning (DAIML) and adaptive control to enable rapid online learning and precise flight control under a wide range of wind conditions. Chapter 5 proposes a lightweight augmentation method that enhances trajectory tracking performance for UAVs by effectively compensating for motor dynamics and digital transport delays. This method is extensible to a range of control methods, including learning-based approaches. Chapter 6 explores a novel sparse failure identification method for detecting and compensating for motor failures in over-actuated UAVs, contributing to the development of robust fault detection and compensation strategies for a safer and more reliable operation. This method builds on the Neural-Fly online learning framework and extends it to handle a wider range of conditions, including complete actuator failures. Together, these chapters address key challenges in safe and reliable learning-based control and demonstrate the potential of deep-learning-based control methods.
", "doi": "10.7907/2xnc-t162", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:15210", "collection": "thesis", "collection_id": "15210", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05262023-141116640", "primary_object_url": { "basename": "HiroDissertation.pdf", "content": "final", "filesize": 18037879, "license": "other", "mime_type": "application/pdf", "url": "/15210/1/HiroDissertation.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Contraction Theory for Robust Learning-Based Control: Toward Aerospace and Robotic Autonomy", "author": [ { "family_name": "Tsukamoto", "given_name": "Hiroyasu", "orcid": "0000-0002-6337-266", "clpid": "Tsukamoto-Hiroyasu" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Pellegrino", "given_name": "Sergio", "orcid": "0000-0001-9373-3278", "clpid": "Pellegrino-S" }, { "family_name": "Doyle", "given_name": "John C.", "orcid": "0000-0002-1828-2486", "clpid": "Doyle-J-C" }, { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Watkins", "given_name": "Michael M.", "clpid": "Watkins-M-M" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "local_group": [ { "literal": "GALCIT" }, { "literal": "div_eng" } ], "abstract": "Machine learning and AI have been used for achieving autonomy in various aerospace and robotic systems. In next-generation research tasks, which could involve highly nonlinear, complicated, and large-scale decision-making problems in safety-critical situations, however, the existing performance guarantees of black-box AI approaches may not be sufficiently powerful. This thesis gives a mathematical overview of contraction theory, with some practical examples drawn from joint projects with NASA JPL, for enjoying formal guarantees of nonlinear control theory even with the use of machine learning-based and data-driven methods. This is not to argue that these methods are always better than conventional approaches, but to provide formal tools to investigate their performance for further discussion, so we can design and operate truly autonomous aerospace and robotic systems safely, robustly, adaptively, and intelligently in real-time.
\r\n\r\nContraction theory is an analytical tool to study differential dynamics of a non-autonomous (i.e., time-varying) nonlinear system under a contraction metric defined with a uniformly positive definite matrix, the existence of which results in a necessary and sufficient characterization of incremental exponential stability of multiple solution trajectories with respect to each other. Its nonlinear stability analysis boils down to finding a suitable contraction metric that satisfies a stability condition expressed as a linear matrix inequality, resulting in many parallels drawn between linear systems theory and contraction theory for nonlinear systems. This yields much-needed safety and stability guarantees for neural network-based control and estimation schemes, without resorting to a more involved method of using uniform asymptotic stability for input-to-state stability. Such distinctive features permit the systematic construction of a contraction metric via convex optimization, thereby obtaining an explicit exponential bound on the distance between a time-varying target trajectory and solution trajectories perturbed externally due to disturbances and learning errors. The first two parts of this thesis are about a theoretical overview of contraction theory and its advantages, with an emphasis on deriving formal robustness and stability guarantees for deep learning-based 1) feedback control, 2) state estimation, 3) motion planning, 4) multi-agent collision avoidance and robust tracking augmentation, 5) adaptive control, 6) neural net-based system identification and control, for nonlinear systems perturbed externally by deterministic and stochastic disturbances. In particular, we provide a detailed review of techniques for finding contraction metrics and associated control and estimation laws using deep neural networks.
\r\n\r\nIn the third part of the thesis, we present several numerical simulations and empirical validation of our proposed approaches to assess the impact of our findings on realizing aerospace and robotic autonomy. We mainly focus on the two joint projects with NASA JPL: 1) Science-Infused Spacecraft Autonomy for Interstellar Object Exploration and 2) Constellation Autonomous Space Technology Demonstration of Orbital Reconfiguration (CASTOR), where we also perform hardware demonstrations of our methods using our thruster-based spacecraft simulators (M-STAR) and in high-conflict, distributed, intelligent UAV swarm reconfiguration with up to 20 UAVs (crazyflies).
", "doi": "10.7907/rznp-g568", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:14994", "collection": "thesis", "collection_id": "14994", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:08052022-231458463", "type": "thesis", "title": "Reliable Learning and Control in Dynamic Environments: Towards Unified Theory and Learned Robotic Agility", "author": [ { "family_name": "Shi", "given_name": "Guanya", "orcid": "0000-0002-9075-3705", "clpid": "Shi-Guanya" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" }, { "family_name": "Yue", "given_name": "Yisong", "orcid": "0000-0001-9127-1989", "clpid": "Yue-Yisong" } ], "thesis_committee": [ { "family_name": "Burdick", "given_name": "Joel Wakeman", "orcid": "0000-0002-3091-540X", "clpid": "Burdick-J-W" }, { "family_name": "Wierman", "given_name": "Adam C.", "orcid": "0000-0002-5923-0199", "clpid": "Wierman-A-C" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" }, { "family_name": "Yue", "given_name": "Yisong", "orcid": "0000-0001-9127-1989", "clpid": "Yue-Yisong" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Recent breathtaking advances in machine learning beckon to their applications in a wide range of real-world autonomous systems. However, for safety-critical settings such as agile robotic control in hazardous environments, we must confront several key challenges before widespread deployment. Most importantly, the learning system must interact with the rest of the autonomous system (e.g., highly nonlinear and non-stationary dynamics) in a way that safeguards against catastrophic failures with formal guarantees. In addition, from both computational and statistical standpoints, the learning system must incorporate prior knowledge for efficiency and generalizability.
\r\n\r\nThis thesis presents progress towards establishing a unified framework that fundamentally connects learning and control. First, Part I motivates the benefit and necessity of such a unified framework by the Neural-Control Family, a family of nonlinear deep-learning-based control methods with not only stability and robustness guarantees but also new capabilities in agile robotic control. Then Part II discusses three unifying interfaces between learning and control: (1) online meta-adaptive control, (2) competitive online optimization and control, and (3) online learning perspectives on model predictive control. All interfaces yield settings that jointly admit both learning-theoretic and control-theoretic guarantees.
", "doi": "10.7907/8rz4-7b35", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:15079", "collection": "thesis", "collection_id": "15079", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:12222022-065507477", "primary_object_url": { "basename": "Tang_Ellande_2023.pdf", "content": "final", "filesize": 128945946, "license": "other", "mime_type": "application/pdf", "url": "/15079/1/Tang_Ellande_2023.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "Studies on Off-Nominal Rotor Aerodynamics for eVTOL Aircraft", "author": [ { "family_name": "Tang", "given_name": "Ellande", "orcid": "0000-0001-5933-4716", "clpid": "Tang-Ellande" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Dabiri", "given_name": "John O.", "orcid": "0000-0002-6722-9008", "clpid": "Dabiri-J-O" }, { "family_name": "Burdick", "given_name": "Joel Wakeman", "orcid": "0000-0002-3091-540X", "clpid": "Burdick-J-W" }, { "family_name": "Tokumaru", "given_name": "Phil", "clpid": "Tokumaru-P" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "As electric Vertical Takeoff and Landing (eVTOL) aircraft become increasingly common, improved understanding of rotor aerodynamics in off-nominal conditions becomes ever more important. A better fundamental understanding of these effects can help inform vehicle design, leading to lower power consumption and improved performance. This thesis will cover a selection of topics to gain a better understanding of the expected rotor aerodynamics associated with use in this class of vehicle, as well as the development of tools to aid in the studies and an analysis of the impact of the effects.
\r\n\r\nTo consider special effects on a rotor in hover on such a vehicle, Chapter 2 is the study of obstructions in the upstream of a propeller, representing the effects of a wing or fuselage blocking a propeller\u2019s inlet. The next is the effect of forward flight on the forces produced by a rotor. Lifting rotors are often used in eVTOL aircraft as the craft transitions to forward flight, so a study of their performance in forward flight as well as a model are presented in Chapter 3. Having examined rotor-wing interactions in hover and isolated rotor performance in forward flight, the next step is to examine rotor-wing interactions in forward flight. Chapter 6 shows the design of an integrated test stand for studying the aerodynamic interactions between lifting propellers and a wing in low-speed, transitional forward flight, as well as the subsequent results.
\r\n\r\nThis thesis also describes the development and implementation of two tools to aid in the work herein. The first (Chapter 4) is a rapid, low-cost method of extracting the geometry of a propeller using photogrammetry which is subsequently used in simulations. The second (Chapter 5) is low-cost and accessible multi-axis force sensor used in the integrated test stand for propeller-wing interaction studies. To assess the impact of the findings, the experimental results and models developed are then taken into consideration by applying them to models of existing eVTOL aircraft in Chapter 7. The change in modeling of hover and transition performance is studied with and without the additional modeling.
", "doi": "10.7907/eytr-nd50", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:15094", "collection": "thesis", "collection_id": "15094", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:01302023-023806052", "primary_object_url": { "basename": "sjhan_thesis-2023-u.pdf", "content": "final", "filesize": 6220970, "license": "other", "mime_type": "application/pdf", "url": "/15094/4/sjhan_thesis-2023-u.pdf", "version": "v6.0.0" }, "type": "thesis", "title": "Control and State-Estimation of Jump Stochastic Systems by Learning Recurrent Spatiotemporal Patterns", "author": [ { "family_name": "Han", "given_name": "SooJean", "orcid": "0000-0003-1195-6465", "clpid": "Han-SooJean" } ], "thesis_advisor": [ { "family_name": "Doyle", "given_name": "John Comstock", "orcid": "0000-0002-1828-2486", "clpid": "Doyle-J-C" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Doyle", "given_name": "John Comstock", "orcid": "0000-0002-1828-2486", "clpid": "Doyle-J-C" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" }, { "family_name": "Wierman", "given_name": "Adam C.", "orcid": "0000-0002-5923-0199", "clpid": "Wierman-A-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis establishes control and estimation architectures that combine both model-based and model-free methods by theoretically characterizing several types of jump stochastic systems (JSSs), i.e., systems with random and repetitive jump phenomena. By expanding the capabilities of model-based stochastic control and estimation, there is potential for artificial intelligence to be implemented as a supplement to theory-influenced design instead of being used end-to-end. We begin by deriving sufficient conditions for stochastic incremental stability for nonlinear systems perturbed by two types of non-Gaussian noise: 1) shot noise processes represented as compound Poisson processes, and 2) finite-measure L\u00e9vy processes constructed as affine combinations of Gaussian white and Poisson shot noise processes. We then present a controller architecture based on a concept we call pattern-learning for prediction (PLP) for discrete-time/discrete-event systems, in which we can take advantage of the fact that the underlying jump process is a sequence of random variables that occurs as repeated patterns of interest. Finally, we demonstrate control and estimation for JSSs in three real-world applications. First, we consider the control of networks with dynamic topology (e.g., power grid with fault-tolerance to downed lines), for which PLP is integrated with variations of the novel system-level synthesis framework for disturbance-rejection. Second, we perform congestion control of vehicle traffic flow over metropolitan intersection networks, for which PLP is extended to pattern-learning with memory and prediction (PLMP) via the inclusion of episodic control, designed to reduce memory consumption by exploiting structural symmetries and temporal repetition in the network. Third, we perform estimation and forecasting (the dual problem to control) for epidemic spread throughout a population network under jump phenomena such as superspreader effects and the emergence of variant viruses. Our results indicate that learning patterns in the jump process makes controller/observer design efficient in data-consumption and computation time, which suggests that it can potentially be used for other JSSs in the real world.", "doi": "10.7907/gyae-jv94", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:14148", "collection": "thesis", "collection_id": "14148", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05142021-163257155", "primary_object_url": { "basename": "nakka_yashwanth_kumar_2021_thesis.pdf", "content": "final", "filesize": 28258884, "license": "other", "mime_type": "application/pdf", "url": "/14148/2/nakka_yashwanth_kumar_2021_thesis.pdf", "version": "v6.0.0" }, "type": "thesis", "title": "Spacecraft Motion Planning and Control under Probabilistic Uncertainty for Coordinated Inspection and Safe Learning", "author": [ { "family_name": "Nakka", "given_name": "Yashwanth Kumar", "orcid": "0000-0001-7897-3644", "clpid": "Nakka-Yashwanth-Kumar" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Burdick", "given_name": "Joel Wakeman", "orcid": "0000-0002-3091-540X", "clpid": "Burdick-J-W" }, { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Yue", "given_name": "Yisong", "orcid": "0000-0001-9127-1989", "clpid": "Yue-Yisong" }, { "family_name": "Hadaegh", "given_name": "Fred", "clpid": "Fred-Hadaegh" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "local_group": [ { "literal": "GALCIT" }, { "literal": "div_eng" } ], "abstract": "During a spacecraft mission design process, engineers often balance the following three criteria: science return, optimality in performance, and safety. Given a science criterion, engineers design the orbit parameters with predefined performance and safety. Often in this approach, the spacecraft has no understanding of the expected outcome or the knowledge of the mission safety criteria. Autonomous science-driven orbit (or goal) selection and planning for safety under uncertainty enable efficient and adaptable missions. To this end, we propose an architecture for information-based guidance and control for coordinated inspection, motion planning and control algorithms for safe and optimal guidance under uncertainty, and architecture for safe exploration.
\r\n\r\nIn the first part of this thesis, we present an architecture for inspection or mapping of a target spacecraft in a low Earth orbit using multiple observer spacecraft. We use an information gain approach to directly consider the trade-off between gathered data and fuel/energy cost. The estimated information gain is a crucial input to the motion planner, which computes orbits and reconfiguration strategies for each of the observers to maximize the information gain from distributed observations of the target spacecraft. The resulting motion trajectories jointly consider observational coverage of the target spacecraft and fuel/energy cost. We validate our architecture in a mission simulation to visually inspect the target spacecraft and on the three degree-of-freedom robotic spacecraft dynamics simulator testbed.
\r\n\r\nIn the second part of the thesis, we present gPC-SCP, Generalized Polynomial Chaos-based Sequential Convex Programming method, to compute a sub-optimal solution for a continuous-time chance-constrained stochastic nonlinear optimal control (SNOC) problem. The approach enables motion planning and control of robotic systems under uncertainty. The proposed method involves two steps. The first step is to derive a deterministic nonlinear optimal control problem (DNOC) with convex constraints that are surrogate to the SNOC by using gPC expansion and the distributionally-robust convex subset of the chance constraints. The second step is to solve the DNOC problem using sequential convex programming (SCP) for trajectory generation and control. We prove that in the unconstrained case, the optimal value of the DNOC converges to that of SNOC asymptotically and that any feasible solution of the constrained DNOC is a feasible solution of the chance-constrained SNOC. We derive a stable stochastic model predictive controller using the gPC-SCP for tracking a potentially unsafe trajectory in the presence of uncertainty. We empirically demonstrate the efficacy of the gPC-SCP method for the following three test cases: 1) collision checking under uncertainty in actuation, 2) collision checking with stochastic obstacles, and 3) safe trajectory tracking under uncertainty in the dynamics and obstacle location by using a receding horizon control approach. We validate the effectiveness of the gPC-SCP method on the robotic spacecraft testbed.
\r\n\r\n\r\nIn the third part of this thesis, we present a new approach for optimal motion planning for safe exploration that integrates the chance-constrained stochastic optimal control with dynamics learning and feedback control. We derive an iterative convex optimization algorithm that solves an Information-cost Stochastic Nonlinear Optimal Control problem (Info-SNOC). The optimization objective encodes control cost for performance and exploration cost for learning, and the safety is incorporated as distributionally robust chance constraints. The dynamics are predicted from a robust regression model that is learned from data. The Info-SNOC algorithm is used to compute a sub-optimal pool of safe motion plans that aid in exploration for learning unknown residual dynamics under safety constraints. A stable feedback controller is used to execute the motion plan and collect data for model learning. We prove the safety of rollout from our exploration method and reduction in uncertainty over epochs, thereby guaranteeing the consistency of our learning method. We validate the effectiveness of Info-SNOC by designing and implementing a pool of safe trajectories for a planar robot. We demonstrate that our approach has a higher success rate in ensuring safety when compared to a deterministic trajectory optimization approach.
", "doi": "10.7907/6329-sf68", "publication_date": "2021", "thesis_type": "phd", "thesis_year": "2021" }, { "id": "thesis:14182", "collection": "thesis", "collection_id": "14182", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05272021-220554457", "primary_object_url": { "basename": "DanielNeamati_SeniorThesis.pdf", "content": "final", "filesize": 7095381, "license": "other", "mime_type": "application/pdf", "url": "/14182/1/DanielNeamati_SeniorThesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "New Method and Analysis of Proximity Trajectory-Only Learned Dynamics for Small Body Gravity Fields", "author": [ { "family_name": "Neamati", "given_name": "Daniel A.", "orcid": "0000-0002-1555-1433", "clpid": "Neamati-Daniel-A" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Minnich", "given_name": "Austin J.", "orcid": "0000-0002-9671-9540", "clpid": "Minnich-A-J" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" }, { "family_name": "Hunt", "given_name": "Melany L.", "orcid": "0000-0001-5592-2334", "clpid": "Hunt-M-L" }, { "family_name": "Ehlmann", "given_name": "Bethany L.", "orcid": "0000-0002-2745-3240", "clpid": "Ehlmann-B-L" } ], "local_group": [ { "literal": "Senior Undergraduate Thesis Prize" }, { "literal": "div_eng" } ], "abstract": "Recent missions to small bodies in the past decade (e.g., Rosetta, Hayabusa 2, and OSIRIS-REx) have reshaped our understanding of small bodies and inspired new, more-capable future missions. Despite the high demand for more missions, large uncertainties in small body properties make missions challenging. Recent work in stochastic optimal control can ensure safety in the face of uncertainty in state, constraints, and dynamics. These stochastic optimal controllers require a model of the underlying dynamics, which is difficult for proximity maneuvers and landing around small bodies. Shape models and finite element-like models are the state-of-the-art for high-fidelity gravity models, but they are computationally expensive and do not readily incorporate onboard data. No gravity model yet exists that can use short-horizon position and acceleration data from recent trajectories onboard in safety-critical autonomous proximity maneuvers and landing. Therefore, we propose a new trajectory-only learning-based method to develop a gravity model. We consider three learning frameworks: Gaussian Process Models, Neural Networks, and Physics-Informed Neural Networks. For each framework, we assess the benefits, computational costs, and limitations of the framework. We found that the Gaussian Process Model generally outperforms the other frameworks in cases of moderate uncertainty. As the uncertainty declines or the data is sufficiently filtered, Neural Networks with spectral normalization provide more accurate gravity models and are computationally cheaper to evaluate. Lastly, we reflect on the methods in this thesis and recommend possible problem reformulations for future research.
", "doi": "10.7907/4csx-4636", "publication_date": "2021", "thesis_type": "senior_major", "thesis_year": "2021" }, { "id": "thesis:14083", "collection": "thesis", "collection_id": "14083", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:02182021-040721884", "primary_object_url": { "basename": "Shi_Xichen_2021_Thesis.pdf", "content": "final", "filesize": 16708714, "license": "other", "mime_type": "application/pdf", "url": "/14083/1/Shi_Xichen_2021_Thesis.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "Intelligent Control for Fixed-Wing eVTOL Aircraft", "author": [ { "family_name": "Shi", "given_name": "Xichen", "orcid": "0000-0002-5366-9256", "clpid": "Shi-Xichen" } ], "thesis_advisor": [ { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Burdick", "given_name": "Joel Wakeman", "orcid": "0000-0002-3091-540X", "clpid": "Burdick-J-W" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" }, { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Yue", "given_name": "Yisong", "orcid": "0000-0001-9127-1989", "clpid": "Yue-Yisong" } ], "local_group": [ { "literal": "GALCIT" }, { "literal": "div_eng" } ], "abstract": "Urban Air Mobility (UAM) holds promise for personal air transportation by deploying \"flying cars\" over cities. As such, fixed-wing electric vertical take-off and landing (eVTOL) aircraft has gained popularity as they can swiftly traverse cluttered areas, while also efficiently covering longer distances. These modes of operation call for an enhanced level of precision, safety, and intelligence for flight control. The hybrid nature of these aircraft poses a unique challenge that stems from complex aerodynamic interactions between wings, rotors, and the environment. Thus accurate estimation of external forces is indispensable for a high performance flight. However, traditional methods that stitch together different control schemes often fall short during hybrid flight modes. On the other hand, learning-based approaches circumvent modeling complexities, but they often lack theoretical guarantees for stability.
\r\n\r\nIn the first part of this thesis, we study the theoretical benefits of these fixed-wing eVTOL aircraft, followed by the derivation of a novel unified control framework. It consists of nonlinear position and attitude controllers using forces and moments as inputs; and control allocation modules that determine desired attitudes and thruster signals. Next, we present a composite adaptation scheme for linear-in-parameter (LiP) dynamics models, which provides accurate realtime estimation for wing and rotor forces based on measurements from a three-dimensional airflow sensor. Then, we introduce a design method to optimize multirotor configuration that ensures a property of robustness against rotor failures.
\r\n\r\nIn the second part of the thesis, we use deep neural networks (DNN) to learn part of unmodeled dynamics of the flight vehicles. Spectral normalization that regulates the Lipschitz constants of the neural network is applied for better generalization outside the training domain. The resultant network is utilized in a nonlinear feedback controller with a contraction mapping update, solving the nonaffine-in-control issue that arises. Next, we formulate general methods for designing and training DNN-based dynamics, controller, and observer. The general framework can theoretically handle any nonlinear dynamics with prior knowledge of its structure. Finally, we establish a delay compensation technique that transforms nominal controllers for an undelayed system into a sample-based predictive controller with numerical integration. The proposed method handles both first-order and transport delays in actuators and balances between numerical accuracy and computational efficiency to guarantee stability under strict hardware limitations.
", "doi": "10.7907/51c6-aa57", "publication_date": "2021", "thesis_type": "phd", "thesis_year": "2021" }, { "id": "thesis:14115", "collection": "thesis", "collection_id": "14115", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:04022021-033321217", "type": "thesis", "title": "Safe and Interpretable Autonomous Systems Design: Behavioral Contracts and Semantic-Based Perception", "author": [ { "family_name": "Cai", "given_name": "Karena Xin", "orcid": "0000-0002-8392-4158", "clpid": "Cai-Karena-Xin" } ], "thesis_advisor": [ { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" } ], "thesis_committee": [ { "family_name": "Burdick", "given_name": "Joel Wakeman", "orcid": "0000-0002-3091-540X", "clpid": "Burdick-J-W" }, { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Chung", "given_name": "Soon-Jo", "orcid": "0000-0002-6657-3907", "clpid": "Chung-Soon-Jo" }, { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "We are on the verge of experiencing a new, integrated society where autonomous vehicles will become a fabric of our everyday lives. And yet, seamless integration of autonomous vehicles into our society will require vehicles to interface safely with humans in an incredibly complex, fast-paced, and dynamic environment. Premature deployment of these new autonomous systems \u2014 without safety guarantees or interpretability of algorithms, could prove catastrophic. How can algorithms governing vehicle behavior be designed in a way that guarantees safety, performance, interpretability and scalability? This is the question this thesis seeks to answer.
\r\n\r\nFirst, we present a framework for architecting the decision-making module of autonomous vehicles so that safety and progress of agents can be formally guaranteed. In particular, all agents are defined to act according to what is termed an assume-guarantee contract, which is broadly defined as a set of behavioral preferences. The first version of the assume-guarantee contract is a behavioral profile, which is a set of ordered rules that agents must use to select actions in a way that is interpretable. With all agents operating according to a behavioral profile, the interactions however, are not necessarily coordinated. We then constrain agent behavior with an additional set of interaction rules. The behavioral profile combined with these additional constraints, are what we term a behavioral protocol. With all agents operating according to a local, decentralized behavioral protocol, we can provide formal proofs of the correctness of agent behavior, i.e. all agents will never collide and agents will make it to their respective destinations. Not only does the protocol so\u00a0defined allow us to make formal guarantees, but it is also designed in a way that scales well in the number of agents and provides interpretability of agent\u00a0behaviors. Safety and progress guarantees are proven and verified in simulation.
\r\n\r\nSecond, we focus on using information from object classifiers to enhance an autonomous vehicle's ability to localize where it is within its environment. The proposed approach for incorporating this semantic information is based on solving the maximum likelihood problem. With a hierarchical formulation, we are not only able to improve upon the accuracy of traditional localization techniques, but we are also able to improve our confidence in the accuracy of object detection classifications. The improvement in robustness and accuracy of these algorithms are shown in simulation.
", "doi": "10.7907/w3m8-es32", "publication_date": "2021", "thesis_type": "phd", "thesis_year": "2021" } ]