Fuzzy logic positionning system

fma38

Hi,

Does anybody has already implemented a fuzzy logic positionning system with a BS2? Is it an appropriate regulation system for a DC motor using a incremental encoder as feedback position?

Thanks,

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Fr

Franklin

Could you describe your application further? What is it you are trying to accomplish?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

fma38

This is the project I started to described in this thread:

http://forums.parallax.com/forums/default.aspx?f=5&m=191601

Instead of using a PID (which will work, for sure), I was wondering if using fuzzy logic could be interesting... and fun! I read articles about fuzzy logic, years ago. But all examples show temperature regulation, with driven heat source. Is a closed-loop system controling a position similar? Or are there fundamental differences, which make fuzzy logic inappropriate in this case?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Fr

Suman

Hello. I have implemented the neural network for positioning of the robotic arm using BS2. But I didnot completely use the BS2 for implementing this neural network. What I did was I took the sensor values into the BS2. Then usin SEROUT I transferred these values to my computer and simultaneouly I executed a .m file in Matlab which accepts these incoming values into the Matlab editor. And then used the NN toolbox in Matlab and the obtained position variables were again tranferred back to BS2. Quite tedious though but successful.

So I guess you can use Fuzzy logic in same way. But I guess Fuzzy Logic is easier to implment in BS2 (than NN). Its matter of writing a bunch of IF loops. But u may run into trouble of floating numbers while dividing the range into vaious membership functions.

And yes implementing a fuzzy controller is quite appropriate in any control problem. The only difficulty in fuzzy control is to better understand the system u r designing, the expected range of input(s) (position as i understand, may be velocity additionally), then taking help of Matlab fuzzy logic toolbox to decide the control strategy and implementing it in hardware.

I neve implemented fuzzy in hardware but in simulations what i found is fuzzy controllers r a bit tough to design but efficient to system disturbances.


Best of Luck

Suman Sadhu

Post Edited (Suman) : 5/28/2007 8:53:18 AM GMT

fma38

Thanks for this feedback. In my case, I do not *really* need fuzzy logic (PID is enough), but it can be fun!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Fr

Tracy Allen

You might find the critical articles by Bob Pease interesting, and entertaining...

www.elecdesign.com/Articles/Index.cfm?ArticleID=4915

Or go to the Electronic Design magazine web site...
www.elecdesign.com/Search/Index.cfm
and type this into the search:
What's All This Fuzzy Logic Stuff, Anyhow?

Or Bob's home page at National Semi...
www.national.com/rap/

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

fma38

Great links! Thanks [noparse]:o[/noparse])

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Fr