Pages: [1]   Go Down
Author Topic: Precise Position Control Of DC Hobby Servo motor Using Fuzzy Logic Controller  (Read 2144 times)
0 Members and 1 Guest are viewing this topic.
Offline Offline
Newbie
*
Karma: 0
Posts: 9
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Hello all,
hopefully everyone is in the best of their health and spirits.

I am required to precisely control the position of a hobby servo motor(HS 311) through fuzzy logic inference system. I require an accuracy upto 0.1 degrees. I have simulated the fuzzy system on matlab and it controls the motor well in time. Now I want to burn this fuzzy inference system in a micro controller.
Can the arduino uno r3 hold the code for this? by 'hold' I mean does it support floating point mathematics to precisely handle the calculations to accurately position the motor.

Any Help or suggestion on the topic will be highly appreciated please.

Thankyou
Logged

Seattle, WA USA
Offline Offline
Brattain Member
*****
Karma: 617
Posts: 49463
Seattle, WA USA
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Quote
by 'hold' I mean does it support floating point mathematics to precisely handle the calculations to accurately position the motor.
Yes, it does. Whether it's fast enough depends.
Logged

0
Offline Offline
Tesla Member
***
Karma: 145
Posts: 9637
Arduino rocks
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Quote
I require an accuracy upto 0.1 degrees.

That is interesting. Using a similar type of hobby servo I got only ~.4 deg of repeatable accuracy.
Logged

Consider the daffodil. And while you're doing that, I'll be over here, looking through your stuff.   smiley-cool

Johannesburg. UTC+2
Online Online
Faraday Member
**
Karma: 99
Posts: 4435
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Is it worth going for such precision with a fuzzy system in the first place?

I know that fuzzy is repeatable, I know how it works, and that it produces a set point based on rules which always give the same answer under the same conditions. (As opposed to a lay understanding which is that the output itself is fuzzy....)

But given that the whole fuzzy membership thing is, well, fuzzy, eg the slopes of the trapezoid ends and stuff like that, and the centroid approach to de-fuzzing, is it really of any value to get too anal about the precision of the setpoint and meeting it in the hardware?

We get to the point where we do anal maths on calculating membership levels and centroid positions (becasue that arithmetic is pretty trivial), losing sight of the fact that the membership definitions of stuff like cold, cool, warm, hot are loose in the first place.
Logged

The Elders of the Internet know who I am
I'm on LinkedIn: http://www.linkedin.com/in/jimbrownza

Pages: [1]   Go Up
Jump to: