I've got karma to burn, as the slashdotters say, so I'll burn it now.

I thought that Fuzzy Logic was done, as in "stick a fork in it," as in do Cox and Kosko and Zadeh even talk about FL any more? I'm rather astonished that someone's still teaching it.

I remember the big fuzzy-logic push back in the early 90s, and I remember our group got the Motorola fuzzy-logic micro development tools.

Fuzzy logic's reason for being is that it provides a higher level of abstraction to controls problems than standard controls theory. And while that's fine for academic papers and press releases, in the Real World where controls live, Fuzzy Logic made the hardware more expensive. Let's see, we can either write traditional PID loop code (simple, concise, well-known) or we can use a tool that translates our nice fuzzy sets into a big bundle of ugly bloated code for your micro. (Yeah, I know that sounds like the same complaint the schematic die-hards used to denigrate VHDL and Verilog, but that war's over and the HDLs have won.)

Fuzzy logic ended up being another buzzword that caught on with the Pointy-Haired Bosses but not with the people who actually implement controls systems.

-a