Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Aeronautics and Astronautics, 2009. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 197-214). / Aerospace systems are among the most complex anthropogenic systems and require large quantities of systems knowledge to design successfully. Within the aerospace industry, an aging workforce places those with the most systems experience near retirement at a time when fewer new programs exist to provide systems experience to the incoming generation of aerospace engineers and leaders. The resulting population will be a set of individuals who by themselves may lack sufficient systems knowledge. It is therefore important to look at teams of aerospace engineers as a new unit of systems knowledge and thinking. By understanding more about how teams engage in collaborative systems thinking (CST), organizations can better determine which types of training and intervention will lead to greater exchanges of systems-level knowledge within teams. Following a broad literature search, the constructs of team traits, technical process, and culture were identified as important for exploring CST. Using the literature and a set of 8 pilot interviews as guidance, 26 case studies (10 full and 16 abbreviated) were conducted to gather empirical data on CST enablers and barriers. These case studies incorporated data from 94 surveys and 65 interviews. From these data, a regression model was developed to identify the five strongest predictors of CST and facilitate validation. Eight additional abbreviated case studies were used to test the model and demonstrate the results are generalizable beyond the initial sample set. To summarize the results, CST teams are differentiable from non-CST teams. / (cont.) Among the most prevalent differentiators is a team's self-reported balance between individual and consensus decision making. Teams that engage in consensus decision making reported stronger engagement in collaborative systems thinking. Another differentiator is the median number of past program experiences on a team. Teams whose members reported more past similar program experiences also reported more engagement in collaborative systems thinking. Data show the number of past similar programs worked is a better predictor than years of industry experience. The apparent enabling effects of qualitative team traits are also discussed. The conclusions of this document propose ways in which these findings may be used to improve training and team intervention within industry, academia, and government. / by Caroline Marie Twomey Lamb. / Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Aeronautics and Astronautics, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 125-126). / The ever-increasing role of intelligence, surveillance, and reconnaissance (ISR) assets in combat may require relatively large numbers of earth observation spacecraft to maintain situational awareness. One way to reduce the cost of such systems is to operate at very low altitudes, thereby minimizing optics size and cost for a given ground resolution. This outside-the-box idea attempts to bridge the gap between high-altitude aerial reconnaissance platforms and traditional LEO satellites. Possible benefits from such a design include enabling a series of cheap, small satellites with improved optical resolution, greater resistance to adversary tracking, and 'quick strike' capability. In this thesis satellite systems design processes and tools are utilized to analyze advanced concepts of low perigee systems and reduce the useful perigee boundary of satellite orbits. The feasibility and utility of such designs are evaluated through the use of the Satellite System Design Tool (SSDT), an integrated approach using models and simulations in MATLAB and Satellite Tool Kit (STK). Finally a potential system design is suggested for a conceptual Continuous Low Orbit Surveillance Satellite (CLOSeSat). The proposed CLOSeSat design utilizes an advanced propulsion system and swooping maneuvers to improve survivability and extend lifetime at operational perigees as low as 160 kilometers, with sustained circular orbits at 240 kilometers. The views expressed in this thesis are those of the author and do not reflect the official policy or position of the United States Air Force, Department of Defense, or the U.S. Government. / by Jared K. Krueger. / S.M.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Aeronautics and Astronautics, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 165-175). / In this thesis, we explore the problem of high-speed autonomous navigation for a dynamic mobile robot in unknown environments. Our objective is to navigate from start to goal in minimum time, given no prior knowledge of the map, nor any explicit knowledge of the environment distribution. Faced with this challenge, most practical receding-horizon navigation methods simply restrict their action choices to the known portions of the map, and ignore the effects that future observations will have on their map knowledge, sacrificing performance as a result. In this thesis, we overcome these limitations by efficiently extending the robot's reasoning into unknown parts of the environment through supervised learning. We predict key contributors to the navigation cost before the relevant portions of the environment have been observed, using training examples from similar planning scenarios of interest. Our first contribution is to develop a model of collision probability to predict the outcomes of actions that extend beyond the perceptual horizon. We use this collision probability model as a data-driven replacement for conventional safety constraints in a receding-horizon planner, resulting in collision-free navigation at speeds up to twice as fast as conventional planners. We make these predictions using a Bayesian approach, leveraging training data for performance in familiar situations, and automatically reverting to safe prior behavior in novel situations for which our model is untrained. Our second contribution is to develop a model of future measurement utility, efficiently enabling information-gathering behaviors that can extend the robot's visibility far into unknown regions of the environment, thereby lengthening the perceptual horizon, resulting in faster navigation even under conventional safety constraints. Our third contribution is to adapt our collision prediction methods to operate on raw camera images, using deep neural networks. By making predictions directly from images, we take advantage of rich appearance-based information well beyond the range to which dense, accurate environment geometry can be reliably estimated. Pairing this neural network with novelty detection and a self-supervised labeling technique, we show that we can deploy our system initially with no training, and it will continually improve with experience and expand the set of environment types with which it is familiar. / by Charles Andrew Richter. / Ph. D.

Massachusetts Institute of Technology. Dept. of Aeronautics and Astronautics. Thesis. 1971. Sc.D. / MICROFICHE COPY ALSO AVAILABLE IN AERO LIBRARY. / Vita. / Bibliography: leaves 441-454. / Sc.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Aeronautics and Astronautics, 1987. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND AERO. / Bibliography: leaves 227-233. / by Thomas Wesley Roberts. / Ph.D.

Thesis: S.M., Massachusetts Institute of Technology, Department of Aeronautics and Astronautics, 2014. / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Cataloged from student-submitted PDF version of thesis. "This is joint work with Jana Tumova, Pratik Chaudhari and Sertac Karaman"--Page 3. / Includes bibliographical references (pages 87-89). / We consider the problem of control strategy synthesis for robots given a set of complex mission specifications, such as "eventually visit region A and then return to a base", "periodically survery regions A and B" or "do not enter region D". We focus on problem instances where there does not exist a strategy that satisfies all the specifications, and we aim to nd strategies that satisfy the most important specifications albeit violating the least important ones. We focus on two particular problem formulations, both of which take as input the mission specifications in the form of Linear Temporal Logic (LTL) formulae. In our first formulation we model the robot as a discrete transition system and each of the specifications has a reward associated with its satisfaction. We propose an algorithm for finding the strategy of maximum cumulative reward which has a significantly better computational complexity than that of a brute-force approach. In our second formulation we model the robot as a continuous dynamical system and the specifications are associated with priorities in such a way that a specification with priority i is infinitely more important than one with priority level j, for any i < j. For this purpose, we introduce a functional that quantifies the level of violation of a motion plan and we design an algorithm for asymptotically computing the control strategy of minimum level of violation among all strategies that guide the robot from an initial state to a goal set. For each of our two formulations we demonstrate the usefulness of our algorithms in possible applications through simulations, and in the case of our second formulation we also carry experiments on a real-time autonomous test-bed. / by Luis I. Reyes Castro. / S.M.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Aeronautics and Astronautics, 2015. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 135-143). / I envision a system that enables successful collaborations between humans and machine learning models by harnessing the relative strength to accomplish what neither can do alone. Machine learning techniques and humans have skills that complement each other - machine learning techniques are good at computation on data at the lowest level of granularity, whereas people are better at abstracting knowledge from their experience, and transferring the knowledge across domains. The goal of this thesis is to develop a framework for human-in-the-loop machine learning that enables people to interact effectively with machine learning models to make better decisions, without requiring in-depth knowledge about machine learning techniques. Many of us interact with machine learning systems everyday. Systems that mine data for product recommendations, for example, are ubiquitous. However these systems compute their output without end-user involvement, and there are typically no life or death consequences in the case the machine learning result is not acceptable to the user. In contrast, domains where decisions can have serious consequences (e.g., emergency response panning, medical decision-making), require the incorporation of human experts' domain knowledge. These systems also must be transparent to earn experts' trust and be adopted in their workflow. The challenge addressed in this thesis is that traditional machine learning systems are not designed to extract domain experts' knowledge from natural workflow, or to provide pathways for the human domain expert to directly interact with the algorithm to interject their knowledge or to better understand the system output. For machine learning systems to make a real-world impact in these important domains, these systems must be able to communicate with highly skilled human experts to leverage their judgment and expertise, and share useful information or patterns from the data. In this thesis, I bridge this gap by building human-in-the-loop machine learning models and systems that compute and communicate machine learning results in ways that are compatible with the human decision-making process, and that can readily incorporate human experts' domain knowledge. I start by building a machine learning model that infers human teams' planning decisions from the structured form of natural language of team meetings. I show that the model can infer a human teams' final plan with 86% accuracy on average. I then design an interpretable machine learning model then "makes sense to humans" by exploring and communicating patterns and structure in data to support human decision-making. Through human subject experiments, I show that this interpretable machine learning model offers statistically significant quantitative improvements in interpretability while preserving clustering performance. Finally, I design a machine learning model that supports transparent interaction with humans without requiring that a user has expert knowledge of machine learning technique. I build a human-in-the-loop machine learning system that incorporates human feedback and communicates its internal states to humans, using an intuitive medium for interaction with the machine learning model. I demonstrate the application of this model for an educational domain in which teachers cluster programming assignments to streamline the grading process. / by Been Kim. / Ph. D.

Thesis: S.M., Massachusetts Institute of Technology, Department of Aeronautics and Astronautics, 2014. / In title on title page, double underscored "x" appears as subscript. Cataloged from PDF version of thesis. / Includes bibliographical references (pages 47-49). / Aviation NOx emissions are byproducts of combustion in the presence of molecular nitrogen. In the upper troposphere, NOx emissions result in the formation of O₃ but also reduce the lifetime of CH4 , causing an indirect reduction in the formation of O₃. Meta studies by Lee et al. and Prather et al. concluded that the short-lived O₃ radiative forcing (RF) was greater than the combined long-lived CH₄ and O₃ RFs, leading to a net positive RF (4.5 to 14.3 mW/m² per Tg of NOx emissions). However, few simulations assess the surface air temperature (SAT) response, or conduct a large ensemble simulation with climate feedback in the cases where SAT is predicted. We aim to quantify the climate forcing and temperature response of aviation NOx emissions. Eight 400-member ensemble simulations are conducted with an earth system model of intermediate complexity. Inter-scenario comparisons between emissions starting in 1991, 2016 and 2036 with mid-range and high anthropogenic emissions are performed. We then determine the existence of long-term temporal heterogeneity of climate forcing and impact. The global net RF of an aviation NO, emissions inventory is positive from 1991 to 2100 while leading to a global average SAT responses of -0.068 K in 2100. Despite the positive zonal RF in the Northern Hemisphere of up to 413.9 mW/m² at 45°N, all latitudes experience cooling after 2075. In another scenario, constant aviation NOx emissions at 4.1 Tg/year cause a global net RF of near zero while leading to a SAT response of -0.020 K in 2100. The unexpected temperature behavior in both scenarios is attributed to the forcing from CH₄ destruction being 64% more effective in generating a SAT response than the O₃ forcing. Despite the positive net RF, the probability of aviation NOx emissions being cooling is 67% because of the relative difference in O₃ and CH₄ efficacies. / by Lawrence Man Kit Wong. / S.M.

Thesis (Sc.D.)--Massachusetts Institute of Technology, Dept. of Aeronautics and Astronautics, 1985. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND AERO. / Vita. / Bibliography: leaves 312-316. / by Steven Ray Hall. / Sc.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Aeronautics and Astronautics, 2004. / Includes bibliographical references (leaves 58-60). / This work presents an approach to diagnosis to meet the challenging demands of modern engineering systems. The proposed approach is an architecture that is both hierarchical and hybrid. The hierarchical dimension of the proposed architecture serves to mitigate the complexity challenges of contemporary engineering systems. The hybrid facet of the architecture tackles the increasing heterogeneity of modern engineering systems. The architecture is presented and realized using a bus representation where various modeling and diagnosis approaches can coexist. The proposed architecture is realized in a simulation environment, the Specification Toolkit and Requirements Methodology (SpecTRM). This research also provides important background information concerning approaches to diagnosis. Approaches to diagnosis are presented, analyzed, and summarized according to their strengths and domains of applicability. Important characteristics that must be considered when developing a diagnostics infrastructure are also presented alongside design guidelines and design implications. Finally, the research presents important topics for further research. / by Martin Ouimet. / S.M.