CaltechTHESIS
  A Caltech Library Service

Modeling, simulation, and design of self-assembling space systems : accurate collision detection, robust time integration, and optimal control

Citation

Johnson, Gwendolyn Brook (2013) Modeling, simulation, and design of self-assembling space systems : accurate collision detection, robust time integration, and optimal control. Dissertation (Ph.D.), California Institute of Technology. http://resolver.caltech.edu/CaltechTHESIS:09132012-125328533

Abstract

Motivated by issues inherent in modeling and designing self-assembling systems (e.g. multiple collisions, collisions between non-smooth bodies, clumping and jamming behaviors, etc.), the goal of this thesis is to develop robust numerical tools that enable ecient and accurate direct simulation of self assembling systems and the application of optimal control methods to this type of system. The systems will be alternately modeled using linear nite elements, rigid bodies, or chains of rigid bodies. To this end, this work begins with development of a linear programming based collision detection algorithm for general convex polyhedral bodies. The resulting linear program has several features which render it extremely useful in determining the force system at the time of contact in numerical collision integrators. With robust collision detection in hand, three related numerical integration methods for dynamics with collisions are treated; a direct potential-based approach, and exact collision integrator in a discrete variational setting, and a decomposition-based algorithm, again in the discrete variational setting. Finally, several control problems are treated in the Discrete Mechanics and Optimal Control{Constrained (DMOCC) framework in which collisions between non-smooth bodies either need to be avoided or explicitly included in the optimal control problem. A globally stable feedback controller and a family of trajectories for spacecraft docking are also developed and tested with an accurate representation of an optimized CubeSat docking system.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:Collision Detection, Linear Programming, Contact, Variational Collision Integration, Optimal Control, Self-Assembly
Degree Grantor:California Institute of Technology
Division:Engineering and Applied Science
Major Option:Aeronautics
Thesis Availability:Public (worldwide access)
Research Advisor(s):
  • Ortiz, Michael
Group:Keck Institute for Space Studies
Thesis Committee:
  • Ravichandran, Guruswami (chair)
  • Ortiz, Michael
  • Leyendecker, Sigrid
  • Pellegrino, Sergio
Defense Date:27 August 2012
Non-Caltech Author Email:gwendolynbrook (AT) gmail.com
Funders:
Funding AgencyGrant Number
Keck Institute for Space StudiesUNSPECIFIED
Record Number:CaltechTHESIS:09132012-125328533
Persistent URL:http://resolver.caltech.edu/CaltechTHESIS:09132012-125328533
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:7203
Collection:CaltechTHESIS
Deposited By: Gwendolyn Johnson
Deposited On:07 May 2013 21:18
Last Modified:04 Mar 2016 21:42

Thesis Files

[img]
Preview
PDF - Final Version
See Usage Policy.

31Mb

Repository Staff Only: item control page