Organizational Unit:
Daniel Guggenheim School of Aerospace Engineering
Daniel Guggenheim School of Aerospace Engineering
Permanent Link
Research Organization Registry ID
Description
Previous Names
Parent Organization
Parent Organization
Organizational Unit
Includes Organization(s)
Organizational Unit
Organizational Unit
Organizational Unit
Organizational Unit
Organizational Unit
ArchiveSpace Name Record
Publication Search Results
Now showing
1 - 10 of 71
-
ItemVerification Of Adversarially Robust Reinforcement Learning Mechanisms In Autonomous Systems(Georgia Institute of Technology, 2022-12-07) Seo, TaehwanArtificial Intelligence (AI) is an effective algorithm for satisfying both optimality and adaptability in autonomous control systems. However, the policy generated from the AI is black-box, and since the algorithm cannot be analyzed in advance, this motivates the performance measurement of the AI model with verification. The performance and safety of the Cyber-Physical System(CPS) are subject to cyberattacks that intend to fail the system in operation or to interrupt the system from learning by modulation of learning data. For the safety and reliability scheme, verifying the impact of attacks on the CPS with the learning system is critical. This thesis proposal focuses on proposing one verification framework of adversarially robust Reinforcement Learning (RL) policy using the software toolkit ‘VERIFAI’, providing robustness measures over adversarial attack perturbations. This allows an algorithm engineer would be equipped with an RL control model verification toolbox that may be used to evaluate the reliability of any given attack mitigation algorithm and the performance of nonlinear control algorithms over their objectives. For this specified work, we developed the attack mitigating RL on nonlinear dynamics by the interconnection of off-policy RL and on-off adversarially robust mechanisms. After that, we connected with the simulation and verification toolkit for testing both the verification framework and integrated algorithm. The simulation experiment of the whole verification process was performed with two different control problems, one is a cart-pole problem from OpenAI gym, and the other problem is the attitude control of Cessna 172 in X-plane 11. From the experiment, we analyzed how the attack-mitigating RL algorithm performed with gain varying specific adversary attacks, and evaluated the generated model performance over the changing environmental parameters.
-
ItemDesign of the VISORS and SWARM-EX Propulsion Systems(Georgia Institute of Technology, 2022-12-01) Hart, Samuel T. ; Lightsey, E. GlennThe Georgia Tech (GT) Space Systems Design Lab (SSDL) will deliver 3-D printed cold gas propulsion systems for the VISORS and SWARM-EX CubeSat formation flying missions. This report provides an overview of the working principle of these and past propulsion systems designed by the SSDL. Further information is provided about the specific designs of each of these systems and the problems encountered throughout the design process. Additionally, recommendations for improvements to future designs are outlined. An analysis of the effects of temperature on these systems is also presented.
-
ItemBest Practices and Considerations for Planning and Conducting Integration of University CubeSats(Georgia Institute of Technology, 2022-12-01) Rawson, WilliamThis paper seeks to serve as a resource for students entering the integration phase of a CubeSat project by compiling best practices and practical considerations from several projects in the Space Systems Design Lab at the Georgia Institute of Technology. The integration phase can be a particular challenge for university CubeSat programs given the value of practical experience in performing these activities and the challenge of managing a student workforce with constant turnover. The topics covered include best practices for planning the integration phase of a project, considerations when performing integration activities, and the characteristics of good assembly procedures. Although the focus is on spacecraft-level integration of CubeSats in a university setting, many of the considerations are applicable outside the academic setting and to subsystem-level integration activities as well. Finally, a case study will be presented illustrating the planning of integration activities for the VISORS mission, a two 6U CubeSat formation-flying mission
-
ItemNumerical Optimization of Satellite Avoidance Maneuvers(Georgia Institute of Technology, 2022-12-01) Sheikh, FirasThis paper investigates the numerical optimization of a collision avoidance maneuver and phase return for a satellite in a constellation network. The time until collision and return time are varied, and three impulse maneuvers are assumed. The total Delta-V is minimized while avoiding the collision by 1 km or greater and returning to the original satellite orbit phase. For this analysis, the Clohessy-Wiltshire (CW) solutions were utilized to linearize the system of equations for relative orbital motion without perturbation assumptions. Results demonstrated that a greater time until collision generally produces minimized Delta-V options but longer total maneuver times as a tradeoff.
-
ItemAn analysis on the application of algebraic geometry in Initial Orbit Determination problems(Georgia Institute of Technology, 2022-12-01) Mancini, MichelaInitial Orbit Determination (IOD) is a classical problem in astrodynamics. The space around Earth is crowded by a great many objects whose orbits are unknown, and the number of space debris is constantly increasing because of break-up events and collisions. Reconstructing the orbit of a body from observations allows us to create catalogs that are used to avoid collisions and program missions for debris removal. Also, comparing the observations of celestial bodies with predictions of their positions made based on our knowledge of the universe has been in the past, and is still today, one of the most effective means to make improvements in our cosmological model. In this work, a purely geometric solution to the angles-only IOD problem is analyzed, and its performance under various scenarios of observations is tested. The problem formulation is based on a re-parameterization of the orbit as a disk quadric, and relating the observations to the unknowns leads to a polynomial system that can be solved using tools from numerical algebraic geometry. This method is time-free and does not require any type of initialization. This makes it unaffected by the problems related to the estimate of the time-of-flight, that usually affects the accuracy of the solution. A similar approach may be used to analyze the performance of the solver when streaks are used, together with lines of sight, as inputs to the problem. Streaks on digital images form, together with the camera location, planes that are tangent to the orbit. This produces two different types of constraints, that can be written as polynomial equations. The accuracy and the robustness of the solver are decreased by the presence of streaks, but they remain a valid input when diversity in the observed directions guarantees the departure from the singular configuration of almost coplanar observations.
-
ItemNumerical Simulations of the HVAB Rotor in Hover(Georgia Institute of Technology, 2022-11-29) Mali, HajarNumerical simulations of compressible viscous flow over the Hover Validation and Acoustic Baseline (HVAB) rotor in hover are presented. The commercial flow solver, ANSYS Fluent, has been employed. The effects of transition are modeled using the Langtry-Menter k-ω SST-γ-Reθ model at three different pitch settings. Comparisons with HVAB test data and other publications are discussed. These include integrated thrust and power coefficients, figure of merit, velocity inflow, surface pressure distribution, lift distribution, transition locations, and vortex structures.
-
ItemCharacterization of Non-volatile Particulate Matter in Pressurized Premixed Laminar Jet-A Flames Via Thermophoretic Sampling(Georgia Institute of Technology, 2022-08-16) Manikandan, Sundar RamProduction and subsequent emissions of non-volatile particulate matter (nvPM) pose a challenge for both optical diagnostics and physical probing, especially at conditions relevant to practical combustors. Key to enabling nvPM mitigation is in-situ optical measurements, particularly laser induced incandescence (LII). However, interpreting the LII signals is challenging. To quantitatively use LII in gas turbines, their measurements must be calibrated and validated against physical nvPM samples. The preferred approach for extracting these physical samples is in-situ thermophoretic soot sampling followed by transmission electron microscope (TEM) imaging. This thesis work deals with the design of a multi-probe thermophoretic soot sampling system capable of extracting nvPM samples in laminar, rich flames of prevaporized jet-A/air premixtures at elevated pressures. The flames under investigation were observed to exhibit thermal-diffusive instabilities, that are responsible for the flame to form corrugated structures. Moreover, these instabilities cause the corrugated flame to exhibit spatio-temporal variations, which exacerbate the challenges in implementing diagnostics. For the soot sampler, a significantly larger sampling time of 125 ms was required to obtain sufficient soot deposition on the TEM grids, which can enhance the extent of restructuring in the deposited soot particles. Visualization of the data through the TEM revealed (i) a wide range of soot particle size varying between 10 – 250 nm; (ii) presence of non-soot organic matters that include (1) fibers, (2) sharp contrasted mineral-like structures, and (3) uniform and porous spherical structures with varying contrast; and (iii) the dominant morphological characteristics of the flame generated soot particles that are indicative of its chemically reactive nature and restructuring. Furthermore, the quantitative results show (i) increasing soot particle size with pressure, and (ii) an increasing-decreasing trend for the mean soot particle size with height above the burner. While the effect of pressure is explained by the enhanced extent of graphitization and maturity in the nanostructures of soot particles at elevated pressures; the dependence on height can be explained through particle agglomeration for the initial increase in size with height, followed by oxidation of the particles respectively. However, considering the range of tested HAB when compared to the flame length, the possibility for inconclusive variation stems from preheat temperature variations and restructuring effects
-
ItemLine of Sight Observational and Inertial Sensing Requirements for Multi-UAV Tracking subject to GNSS Denial(Georgia Institute of Technology, 2022-08) Bouhall, Matthew M.
-
ItemImpacts of Altitude on Expectation Maximization-based Data Association and Orbit Estimation(Georgia Institute of Technology, 2022-08) Sakson, Daniel
-
ItemIntegration and Testing of a 2U Cold-Gas Propulsion System for the SunRISE Mission(Georgia Institute of Technology, 2022-08-01) Shirazi, Kian E.The Georgia Tech (GT) Space Systems Design Laboratory (SSDL) is building six identical cold-gas propulsion systems to provide the necessary maneuvering capabilities required by the SunRISE mission. The mission plans to observe low frequency emissions from the Sun by utilizing an array of CubeSats that will formation fly to create a large radio telescope in space. The cold-gas system design is based on the lab’s heritage design used for previous missions, namely BioSentinel, which leverages additive manufacturing to create a highly optimized propulsion system for the CubeSat form factor. The unique system consists of mainly a singular printed multifunctional structure encompassing tanks, plumbing, and nozzles, and utilizes a two-phase propellant to maximize the amount of propellant stored in the restricted volume and hence the total impulse provided by the system. This report provides a brief overview of the system design and its purpose in the SunRISE mission, while detailing the integration process and extensive testing campaign each flight unit goes through before they are delivered for integration with the spacecraft.