23 Apr 06 Meeting Notes

From FBSwiki
Revision as of 17:33, 8 July 2006 by Murray (Talk | contribs)

Jump to: navigation, search

Meeting with Karl in Cambridge, UK

General discussion

  • Robust performance section: focus on disturbances, noise rejection
  • New section on noise (3-5 pages)
    • Karl to take first cut (done)
    • Distribute time white noise, spectral density
  • Move implementation chapter to wiki (done)
    • Action (RMM): Re-insert bio stuff into wiki version of implementation chapter


Chapter 1 discussion

  • Thoughts on controls tools (chapter 1)
    • What we consider a solution changes with modern computation
    • Theme throughout: standard problems/models → computational solutions
    • Examples of standard problems: ABCD, block diagrams, transfer functions, LQG, Hinfty, MPC/RHC
    • Need an "architecture for solving problems"
    • NCS drawing (from CDS 110b)
  • Action (KJA): Make sure to mention modes and Heaviside

Diagram Figures

  • Circuit: PID or automatic gain control
  • Gas turnbin OK
  • Bio from mCSHL
  • Chemical process example
  • Petri net

References

  • Advice for where to learn more
  • Specialized books
  • Historical perspective

Reachable canonical form

  • Action (RMM): check sources for canonical forms: Kalman, Ho Narenda; Elmer Gilbert; Kalmbin, Arbib, Falb
  • Decision: use the form with the coefficients in order (w/ a negative sign) in the top row
    • Closed loop polynomial should be

LQR/LQG

  • 's for controller cost
  • 's for noise covariances
  • Action (RMM): Update LQR section to reflect this notation
    Completed on 18 May in text and summary

State feedback notation

  • Final results should be A - BK - LC
  • This gives
  • Action (RMM): This is not consistent; figure out what we really want