Society of Robots - Robot Forum

Electronics => Electronics => Topic started by: king28 on January 21, 2009, 02:09:04 AM

Title: Linear Quadratic Regulator to DESIGN & DEVELOPMENT OF A SELF BALANCING ROBOT
Post by: king28 on January 21, 2009, 02:09:04 AM
Hi, everybody , i am great to discover this forum.

I am now l doing my project  using Linear Quadratic Regulator to  DESIGN & DEVELOPMENT OF A SELF BALANCING ROBOT, and stimulate it by using MAtlab 7.5.

Any one have experienced on it?

 now i am looking for the suitable devices and material  , any recommendation ? Some people suggest that using MC68HC12 is that works?Or we should use CMD-555?



Any recommendation  from u all guys??about the circuit, devices or whatever...hope can post it here to help me .

I will appreciate ur help.


Thank you
Title: Re: Linear Quadratic Regulator to DESIGN & DEVELOPMENT OF A SELF BALANCING ROBOT
Post by: cosminprund on January 21, 2009, 03:07:17 AM
Doing an balancing bot seems to be an very complex task that requires a lot of research. It has been done before and you can find some references but (to my limited knowledge) there's no tutorial on how to do real balancing.

You can see an tutorial on how to simulate balancing here:
http://www.societyofrobots.com/member_tutorials/node/185
In my opinion that's not balancing, that's keeping the base parallel to the surface - it only looks like balancing, but it's a start.

It seems that the most famous balancing bot is the "nBot", read up on that, you'll get lots of links to other balancing bots once you get to the nBot page.

You should probably do a bit more research, figure out exactly what you want to do and then come back with better questions, like: "Hey, I'm using this 3D accelerometer and this gyro and I'm not sure what MCU to use. Since the 3D accelerometer can only be read at about 300Hz and the gyro about 1Khz, is this MCU fast enough?". The idea is to do your own research and ask specific questions, giving as much information as possible.

Hope this helps.

P.S: Don't double-post, that never helps.
Title: Re: Linear Quadratic Regulator to DESIGN & DEVELOPMENT OF A SELF BALANCING ROBOT
Post by: Soeren on January 21, 2009, 07:42:32 AM
In my opinion that's not balancing, that's keeping the base parallel to the surface - it only looks like balancing, but it's a start.
Well, it's OK to disagree with the established science world, but a rose is still a rose to the rest of us  :P
Would you claim that a helicopter isn't flying, since it uses other methods than a regular aeroplane?

Balancing is balancing, whether your means of sensing an off-balance state is done by U.S., accelerometer, gyro, trained monkey or even your dishwasher/SO.

Apart from that, I agree that king28 should do a bit of research him-/herself before asking educated questions, or he will loose all rights to state "I am now l doing my project ", as it would neither be "I am doing", nor "my project" anymore, but should be changed to "I got someone to do it for me, fear me if I graduate".
Title: Re: Linear Quadratic Regulator to DESIGN & DEVELOPMENT OF A SELF BALANCING ROBOT
Post by: cosminprund on January 21, 2009, 08:59:11 AM
In my opinion that's not balancing, that's keeping the base parallel to the surface - it only looks like balancing, but it's a start.
Well, it's OK to disagree with the established science world, but a rose is still a rose to the rest of us  :P
Would you claim that a helicopter isn't flying, since it uses other methods than a regular aeroplane?

:) Well I guess it boils down to how one define "balance". Since English is not my first language I asked google for a bit of help. Google suggested those two:
http://en.wikipedia.org/wiki/Balance_(ability) (http://en.wikipedia.org/wiki/Balance_(ability))
http://www.thefreedictionary.com/balancing

And
The Wikipedia definition links "balancing" to the center of gravity and by that definition the Ro-Bot-X is not a balancing robot (because it maintains an fixed angle between itself and the surface it lays on, irrelevant of how this moves it's center of gravity in relation to it's base of support).

The second definition (and that's what's probably the more usual one) links balance to "canceling out forces" - that sounds allot more scientific :D By that definition the Ro-Bot-X is an balancing robot... unless it has to move across a surface that changes inclination.

And yeh, trained monkeys are good at keeping balance :P
http://www.youtube.com/watch?v=Pc2GJLvfuOY
http://www.youtube.com/watch?v=s-EHDtfYazw&feature=related
Title: Re: Linear Quadratic Regulator to DESIGN & DEVELOPMENT OF A SELF BALANCING ROBOT
Post by: aludra_55 on April 20, 2009, 07:05:07 PM
hi... i got this monkey to balance... it looks like a robot though...

made up of 1 axis accelerometer and no gyroscope.. no encoders either...
Title: Re: Linear Quadratic Regulator to DESIGN & DEVELOPMENT OF A SELF BALANCING ROBOT
Post by: msprague on April 21, 2009, 06:26:29 AM
You might like to read this as well:
http://www.wa4dsy.net/robot/balancing-robot/analog-balancing-bot (http://www.wa4dsy.net/robot/balancing-robot/analog-balancing-bot)
Dale built a balancing bot without a microcontroller. There is no obstacle detection or navigation, but it does the balancing very well. He has a very nice writeup on it.