11 Comments
User's avatar
Maxim Raginsky's avatar

FWIW, whenever I teach undergrad control, I always aim to emphasize one key fact: complex exponentials of the form exp(st), where s is complex, are eigenfunctions of linear time-invariant systems. The reason why you need s to be complex (as opposed to purely imaginary, which would only give you sinusoids) is that the corresponding set of signals (sinusoids with exponentially decaying or exponentially growing envelopes) is sufficiently rich to allow for things like system ID and for analyzing both transient and steady-state behavior.

Expand full comment
Ben Recht's avatar

Yeah, but the mere fact that exp(st) are eigenfunctions of any linear system is mind blowing. I know the math is simple, but it's a crazy fact, man.

And that this "linearity" mathematics applies all over the place is one of the biggest technological miracles of the 20th century.

Expand full comment
Cagatay Candan's avatar

I try to focus only on the delay to make things simple. A delayed sinusoid (real or complex valued) is another sinusoid with the same frequency. Delay can be large or small (even infinitesimal) does not matter and the delayed and sum sinusoids interfere constructively, destructively and ... (eventually everyone is asleep!)

Expand full comment
Door Hambilgorb's avatar

Loving this series. Has me side-eyeing the spine of my Ogata textbooks from across the room.

Note sure if this fits in with your gameplan yet, but I'd be interested to hear how/if robust control fits into the modern optimization-industrial complex.

Expand full comment
Ben Recht's avatar

Thanks! And yes, the rest of the week is about robust control and whether any of it mattered/matters. Stay tuned!

Expand full comment
Joe's avatar

Prof. Recht -- killer blog, love literally everything you post. Looks like you have some conversion problems at some point in the Mathjax/Latex > Markdown > HTML > live site pipeline, though -- I see raw Latex in the post. Or at least, that's my guess; my blog (the less interesting social science grad student version of this: (https://griffinjmbur.github.io//counting-blogpost/) suffers from similar problems. I write in Mathjax and only one kind of in-text Latex escape character works for getting it to a GH blog with Jekyll (the clunky \\(x+y\\) style), which I did not realize until I began publishing my notes, so I've made a lot of manual conversions from $x+y$ to that (unfortunately ChatGPT is strangely bad at automating this), and I have lots of stuff like "\(\frac{P}{1+PC}\)", which I see in much of your post, in there as a result.

Just a friendly heads-up. You are literally my favorite blogger about anything remotely serious.

Expand full comment
Ben Recht's avatar

Hey Joe, Thank you so much for the kind words.

With regards to mathjax, it's sadly a Substack problem. If you refresh the page, the TeX will render correctly. I'm not sure if there's a good fix here, but perhaps I should file a support ticket... I'm probably the only person on here who's using the feature.

In future posts, I might just use LaTeXit and export pngs. Clunkier, but always rendered correctly.

Expand full comment
FourierBot's avatar

robotics are powerful

Expand full comment
Oluwasegun Somefun's avatar

A whole new world when we replace Laplace with z-transform: z = exp(st).

Expand full comment
Damek Davis's avatar

> The tricky complex analysis sticks when you hear a concept in a song you like.

Missing an opportunity to include audio examples!

Expand full comment
Biomeg's avatar

Interesting, I never thought about Laplace transforms for sound. I have used them in EEG signal source decoding, and we tend to call it the Laplacian transform. Math is everywhere!

Expand full comment