L2: Requirements

Lecture 2: Requirements

Lecture 2 slides Download Lecture 2 slides

Requirements are fundamental, and should always guide control design.

This lecture provides the tools for discussing and quantifying the behavior and properties of feedback systems.

Preparations

Remind, or familiarize, yourself with

  • Analytic/Holomorphic functions
  • Complex curve integration: Cauchy's theorem, residues, residue theorem
  • Parseval's and Placherel's theorems
  • Maximum modulus theorem

Here are some resources:

Complex analysis cheat sheet Links to an external site.

Sven Spannes blixtkurs i komplex integration Links to an external site.

Tank explorations

The tank experiments were made in Julia with Moberg.jl interfacing directly with the tanks. The scripts are here (warning Messy)

tank_setup.jl Download tank_setup.jl Collect static info

tank_noise.jl Download tank_noise.jl Collect noise info

empirical_tfs.jl Download empirical_tfs.jl Naive frequency domain experiments

test_plots.jl Download test_plots.jl Statics plots

tank_noise_plot.jl Download tank_noise_plot.jl Plot noisy data and compute RMS

empirical_tfs_plots.jl Download empirical_tfs_plots.jl Plot empirical transfer functions

 

Additional Reading

The lecture covers the material

Chapters 2, 10.3, 12.1, 12.2, 12.4, 13.1--13.3 in Feedback systems, Åström Murray Links to an external site.

The proof of the approximate inverse theorem is from KJ 1968 page 241. Note the mistake on the 9th line from the bottom: it's missing a factor two coming from the symmetric derivative

lims0f(s)f(s)s=2f(0)

Chapter 8 in Boyd & Barratt Links to an external site.collects mathematical definitions of common (and quite a few esoteric) closed-loop performance specifications. Chapters 9 and 10 discuss robustness specifications (the role of S and T in differential sensitivities) and robustness via gain-bounds.