Fuzzy control is not suitable for everything but it does some things
very well.
I became interested in it when I started experimenting in autopilots for
sailboats. (Power boats tend to be much easier)
There is some really horrible autopilot software out there. Its amazing
how bad some of it is.
There are several articles out there on Boat, Ship and airplane
Autopilots being implemented with fuzzy logic.
Dave
On 10/19/2018 11:20 PM, Gene Heskett wrote:
On Friday 19 October 2018 22:36:43 Dave Cole wrote:
Here is one I remember.
http://www.rentanadviser.com/en/pid-fuzzy-logic/pid-fuzzy-logic.aspx
There is also a lot of developed open source Fuzzy logic libraries.
Do a search for "Fuzzy Logic Open Source".
Like everything else, some of it is good, some of it is garbage.
Faster than a PID? I don't know if it can be faster, but done right,
I think it can be more effective than a PID. Especially if you really
need something more sophisticated than a simple PID loop.
I haven't so far.
Normally for a PID, you will have a PV(Process variable) and a
Setpoint for the PV and math is run that creates an Output that
affects the PV, hence the "loop".
Looking at the math in the first goggle hit, lots of it, although its all
integer, seems like it would have to be slower.
A PID, since I put the new encoder on the G0704, is good enough that my
only clue its overloaded, is the motor's chirp from Jons pwm-servo amp
as it goes into current limiting at about 17 amps. Up to that point, no
detectable slowdown. I figure I'm pulling around 2hp at that point, its
a 1hp 90 volt motor with a nameplate amps in the middle 9's, and its
being hammered with 126 volts and nearly double the amps. And working
like it was made for it. Knocking on wood, even the plastic gears in the
2 speed head are holding up well. I've even put in some extra hal
trickery for a seemless gear change while its turning, as the tally
switches also are tied into the drive so that if both are open, the
motor revs drop to about 30 while the shifter is moving, and returns to
the set speed as the knob hits the last 2 degrees of the stop closeing
one of the switches. With that encoder and the PID setup, the speed
changes are done in a few milliseconds, way faster than I can muscle the
knob. That also is tied to the tach, so the tach is accurate in either
gear.
But I'm always curious about what may be a better way,
Thanks Dave.
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
