Meeko Mitsuko K. Oishi

Harry S. Truman Postdoctoral Fellow in National Security Science and Engineering
Sandia National Laboratories

Ph.D. (Mechanical Engineering, Design Division) Stanford University, 1/2004, under the supervision of Professor Claire Tomlin, Department of Aeronautics and Astronautics.
M.S. in M.E. (Mechanical Engineering, Design Division) Stanford University, 3/2000.
B.S.E. in M.E. (Mechanical and Aerospace Engineering) Princeton University, 6/1998.

Research

• Safety-Critical Control for Power Systems
  In collaboration with researchers in the National Infrastructure Simulation and Analysis Center at Sandia National Laboratories, I am examining how to design safety-preserving control laws for large, interconnected systems. Addressing systems of such high continuous dimension is made possible by breaking the problem into separate, decentralized calculations. The decentralized reachability calculations determine a conservatively ``safe'' operational space that assumes a neighboring node's state is unpredictable. The computed result will provide a subset of the states from which safe operation is guaranteed.
• Modeling Social Dynamics in Animal Groups
  In my current work, a joint project with Joan Roughgarden, Biological Sciences, we are using cooperative game theory to model social dynamics in animal groups. The goal of this work is to use mathematical modeling to better understand coalition formation and social structures not adequately explained by traditional Darwinian theory.
• User-Interfaces for Hybrid Systems (Abstract, Thesis)
  In complex, hybrid systems, such as civil jet aircraft, information about the system must be carefully chosen for display to the human controlling the system. Pilots can be overwhelmed by excessive information about the aircraft, yet with too little information, a pilot risks not being able to safely control the aircraft. My thesis research focused on 1) modeling human interaction with a hybrid dynamical system, and on 2) creating succinct representations of hybrid systems with which a human can interact. I developed an abstraction technique, applied it to the verification of information shown on a pilot's display of a large civil jet aircraft, and validated my results on a modern commercial aircraft flight simulator.
• Discrete Observability of User-Interfaces
  ``Good'' user-interfaces have certain observability properties, which allow for determination of the state based only on current information (as opposed to a history of observations). We called this type of observability ``immediate observability'', and used it to derive conditions for output synthesis. We applied the output synthesis to the problem of minimizing information broadcast to a fleet of formation-flying aircraft.
• Flight Management Systems
  We model the longitudinal dynamics of a VSTOL aircraft in three modes of flight (conventional, vertical, and transitional) and study the hybrid stability of the switched system. Flight envelope constraints within each flight mode generate generate different ``safe'' regions of operation within each flight mode. Creating a controller to accomplish both objectives (safety and stability) is a non-trivial problem, and a current area of work. In previous work, in collaboration with researchers at Honeywell Technology Center, we calculated optimal trajectories within a given flight envelope using nonlinear constrained optimization.

Awards

• Truman Fellowship in National Security Science and Engineering, Sandia National Laboratories, 2005-2006.
• National Science Foundation Graduate Research Fellowship, 1998-2000, 2001-2002.
• John Biewnkowski Memorial Prize, Mechanical and Aerospace Engineering, Princeton University, 1998.

Publications

Internships

• National Academy of Sciences, 2004. Science and Technology Policy Graduate Fellow, Board on Mathematical Sciences and Their Applications, Washington, D.C.
• Honeywell Technology Center, 2000. Guidance and Control group, Minneapolis, Minnesota.
• Boeing, 1998. 767 Electrical Equipment Installation, Seattle, Washington.
• Intel, 1997. Environmental Engineering, and Automation, Albuquerque, New Mexico.
• Sandia National Labs, 1994, 1995, 1996. Computer Architectures (1996, 1995), Computer Networking (1994), Albuquerque, New Mexico.

Background Information

Meeko Oishi (moishi at stanfordalumni dot org)

Last updated 1/29/06.