| Event Details |
|---|
|
Thursday, February 5, 2015 Talk:
4:00 p.m., Avery 115
Reception:
3:30 p.m., Avery 348
|
Dr. Justin Bradley
Research Fellow, University of MichiganAbstract
Modern vehicles are equipped with a complex suite of computing (cyber) and electromechanical (physical) systems. Holistic design, modeling, and optimization of such Cyber-Physical Systems (CPS) requires new techniques capable of integrated analysis across the full CPS. We introduce a mechanism for balancing cyber and physical resources in a step toward holistic co-design of CPS. First, an ordinary differential equation model abstraction of controller sampling rate is developed and added to the equations of motion of a physical system to form a holistic discrete-time-varying linear system representing the CPS controller. Using feedback control, this cyber effector, sampling rate, is then co-regulated alongside physical effectors in response to physical system tracking error. This technique is applied to attitude control of a small satellite (CubeSat). Additionally, two new controllers for discrete-time-varying systems are introduced; a gain-scheduled discrete-time linear regulator (DLQR) in which DLQR gains are scheduled over time-varying sampling rates, and a forward-propagation Riccati-based (FPRB) controller. The FPRB CPS controller shows promise in balancing cyber and physical resources.
Speaker Bio
Justin Bradley is a Research Fellow in the Department of Aerospace Engineering at the University of Michigan. Justin holds a B.S. in Computer Engineering and M.S. in Electrical Engineering from Brigham Young University, and M.S. and Ph.D. degrees in Aerospace Engineering from the University of Michigan. He spent several years working on the National Ignition Facility at Lawrence Livermore National Laboratory as a Control Systems Software Engineer. He has worked with UAS for over 7 years as a member of the MAGICC lab at BYU, and the A2Sys lab at the University of Michigan. His current research focuses on Cyber-Physical System co-design techniques primarily in the context of aerospace and robotics applications.