0
$\begingroup$

I am trying to design a control system to move a joint of four bar linkage according to desired trajectory. A dc-motor is used to move the joint. Can fuzzy logic controller be used to solve this problem?

I have already implemented a PID controller to do this task. I want to explore other control strategies to do this

Improve this question
$\endgroup$
3
  • $\begingroup$ Do you want to promote Fuzzy logic as a problem solving technique? $\endgroup$ – Manuel Rodriguez Aug 23 '19 at 17:38
  • $\begingroup$ No. I want to see if i can get better performance (minimize error) in trajectory tracking $\endgroup$ – vacky Aug 23 '19 at 19:03
  • 1
    $\begingroup$ You can, however if you want to get into nonlinear systems, it may be more relevant to look into other nonlinear controllers. Siding mode for example or optimal trajectory and linear schedulers...be warned, the fantastic linear techniques of PID or LQR or pole placement don’t apply so directly anymore...the increase in complication is rarely worth the extra effort. $\endgroup$ – morbo Aug 24 '19 at 13:43
1
$\begingroup$

Yes, you can use any kind of algorithm to control a system.

However, I wouldn't suggest preferring a fuzzy logic based controller over PID just because it yields better performance! Because it is hard to prove, if not possible, the stability of the closed-loop system when there is a fuzzy logic controller in the loop. That is the reason why the fuzzy logic has not been used in major control applications such as space, aviation, safety-critical systems, etc.

Just a friendly advice: you can study fuzzy logic for your academic curiosity, but don't expect it will shine on your resume when looking for controls engineering jobs.

Improve this answer
$\endgroup$
6
  • $\begingroup$ It is a fallcy to think that fuzzy logic/controllers can only be delegated to academic persuits, see this robotics.stackexchange.com/a/18886/23148 answer. $\endgroup$ – morbo Aug 24 '19 at 13:36
  • $\begingroup$ @morbo Let's agree to disagree. Fuzzy logic does not bring any additional capability over the existing methods. That is, all control problems that can be solved by fuzzy logic methods can also be solved by classical control methods. It is easier to prove stability, performance of a system when a classical controller is used. That is why fuzzy logic was not implemented in industry applications. I haven't seen any flight control software, self-driving car control software, robotics software that use fuzzy controller. $\endgroup$ – Octavius Aug 24 '19 at 17:35
  • $\begingroup$ I agree its not a solution brought down by god, it never was and was marketed poorly. I also never mentioned the points you suggested. It however does solve and has solved many issues that classic methods simply have not, as you didnt read my links, your statements are closer to dogma than fact (of which your statements are neither) You may have not seen such things, however others have, including my self and many many other researchers. I agree its not the best. I don’t agree with you. $\endgroup$ – morbo Aug 24 '19 at 18:27
  • $\begingroup$ @morbo I don't mind (care) if you don't agree with me. My statements are based on my work experience both in academia and industry not dogma! Good luck with your research in fuzzy logic. $\endgroup$ – Octavius Aug 24 '19 at 19:13
  • $\begingroup$ I don’t research in fuzzy. But clearly nobody else on earth needs to either, since your work experience has closed the subject. $\endgroup$ – morbo Aug 24 '19 at 19:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.