Using the Bees Algorithm to tune a fuzzy logic controller for a robot gymnast
This paper focuses on using the Bees Algorithm to tune the parameters of a fuzzy logic controller developed to stabilise and balance an under-actuated two-link acrobatic robot (ACROBOT) in the upright position. A linear quadratic regulator (LQR) was first developed to obtain the scaling gains needed to design the fuzzy logic controller. Simulation results confirmed that using the Bees Algorithm to optimise the membership functions and the scaling gains of the fuzzy system improved the controller performance.
| Attachment | Size |
|---|---|
| Gymnast1.wmv | 2.92 MB |
Maybe, Japanese wasps have attacked two papers of the bee’s session and killed all the collected honey.
We need to protect our hive from strangers.
However, bees are active enough and can make new honey.
Sameh Otri
PhD Research Student
Hi Sameh,
Our Cardiff Bay Bees are peace-loving and industrious creatures.
They work hard to gather nectar and make honey for all of us.
They have NO enemies at all.
The file deletions were mistakes - probably committed by one of the Bees.
There was no deliberate attempt, as far as we know, by any wasps to steal honey from the hive nor to harm any of our Bees!!!
Best wishes.
DTP.
PS: In any case, wasps, bears, humans, etc. are all welcome to share our honey with us, i.e. use our Bees Algorithm in their work as long as they remember to cite the source!
Dear authors,
Very nice paper and I am so keen to watch the presentation, but seems there is a problem with voice of the presentation, I can not hear any voice with the presentation. Would you please check it out.
Thanks,
Afshin
You are absolutely right.
We are pleased to see you around us,
As you always support us,
Regards.
Otri
come on Afshin where are your qestions?
Bees are ready for all type of questions....
Looking forward to seeing your questions....
Thanks
Sameh Otri
PhD Research Student
Dear authors,
The intial patch size seems to be very small, ngh=0.025. If you are using this initial size, what is the patch size at the end, and what is the range of parameters to search(search space)?
Regards,
Afshin
The patch size is fixed.
Regarding the range of serach space, it is different,
each one of the parameters has its search space, they are available in the paper.
Is there any research done using GA or any other Intelligent Optimisation techniques?
and if any , why it is not included in your rsults?
Thanks
Sameh Otri
PhD Research Student
Yes, there were some works done with GA, but not with the same controller so they are not comparable. This paper is to present the capabilities of the Bees in Optimising fuzzy systems and its use in robotics applications
Assalamualaikum Ahmed and Hai Prof. Pham.
Hai, Ahmed. How do you find the patch size (0.025). Have you tried with various value of the patch and what are the results?
Regards,
Shah
Hello Dr Shah!
Good to hear from you.
We hope you are beesily working on forming your colony back home!
We look forward to receiving your contribution to IPROMS 2008.
Best wishes to you and your family from all the Bay Bees.
DTP.
PS: I shall leave it to Ahmed to reply to your question. As you know, I am not a believer in too small an initial patch size.
Hello Dr Shah
Regarding the patch size, i used this small value because of the small search space of the parameters of the membership functions.
In this paper I used fixed value for the patch size, and I will use different patch sizes for each dimension in future work.
Actually I tried many values for the patch size and all of them gave good results, I didn,t mentioned them because of lack of space in the paper, where each value of the parameter should be illustrated in figures.
So I took only one result and put it in the paper.
Hello Shah Bee
Good to hear from you and hope your fine up there.
Dear authors, I have a suggestion about patch size, what if you use i.e. 0.1 for the intial patch size and reduce the size to 0.025 in some stages during the running of the programe, or use the shrinking method described in the modified version of the Bees Algorithm ( Multi-Objective function Optimisation using the Bees Algorithm) to reduce the patch size automatically and regarding the performance of the programe.
Regards,
Afshin
Thanks for your suggestion, Afshin, about patch size, which we will consider in our future work.
Sameh
Wa alukom assalam Dr Shah-Bee,
Nice to see you around.
any new Bees idea?
Take care and keep in touch
Best wishes
Sameh Otri
PhD Research Student
Hello Bay Bees,
Where are the comments gone?
Where are the Bay Bees gone?
Hopefully not too far away!
Cheers.