2 axis gyro modules

[daveski]

Is there going to be a 2 axis Gyro modual for the Bioloid?
As this would be a factor that would sway me to buy the Bioloid kit I wanted to know if anyone knows anything about the statment that I found at the link below. If you have any more info please post it here.

According to the Robotis staff at their booth, the CM-5 can also interface with 2 axis gyro modules

I have emailed Robotis about this statment and as yet I am still awaiting a response.

Also, does anyone know when the expert kit will be out and how much it will cost?

Thanks

Daveski

[JonHylands]

I'm in the process of building a 5 and 6-axis IMU for the Bioloid. You can see details here:

http://www.bioloid.info/tiki/tiki-index.php?page=5-Axis+Bus+IMU
http://www.bioloid.info/tiki/tiki-index.php?page=6-Axis+Bus+IMU

I don't know when it will be done, but I have my first prototype AVR sensor board wired up and actually running some code, so hopefully within the next two or three weeks I will be sending away to have these boards printed.

Depending on demand, I may have a bunch of them printed. The two sensor boards for it cost $175 from Sparkfun. The 5-axis version will be $65 cheaper (you only need the $110 5-axis board).

- Jon

[daveski]

I'm in the process of building a 5 and 6-axis IMU for the Bioloid. You can see details here:

http://www.bioloid.info/tiki/tiki-index.php?page=5-Axis+Bus+IMU
http://www.bioloid.info/tiki/tiki-index.php?page=6-Axis+Bus+IMU

I don't know when it will be done, but I have my first prototype AVR sensor board wired up and actually running some code, so hopefully within the next two or three weeks I will be sending away to have these boards printed.

Depending on demand, I may have a bunch of them printed. The two sensor boards for it cost $175 from Sparkfun. The 5-axis version will be $65 cheaper (you only need the $110 5-axis board).

- Jon

Wow, looking good. I'm sure there will be a lot of interest in any Inertial Measurement Unit that can be used with the Bioloid.

Check this vid btw, its not using an IMU but just the current load on the ankle servos.

http://www.youtube.com/watch?v=Q8bMvv9iSu4

There are a few more vids and will post them as soon as I can find them.


The servo current load method can be used for static balancing e.g. standing on a slope or uneven ground and a gyro for dynamic balance, e.g. accelerative forces.

Correct me if I'm wrong but IMUs contain Piezo crystals that only detect Acceleration and so hence can't tell if the bot is on a slope unless it starts to topple by which time it would probably be too late.

Anyway, if you do produce these IMUs then put me down for 1. I'm sure with static and dynamic balance the bioloid will be a an awsom bot.

Thanks for the reply.

Daveski

[JonHylands]

I've got some graphs of a six-axis IMU on my Bioloid MicroRaptor:

http://www.bioloid.info/tiki/tiki-index.php?page=MicroRaptor+Screenshots

You can see the accelerometers react very well to falling over. Whether or not they can sense it fast enough to prevent the fall is another question, one which I hope to answer in a few weeks...

- Jon

[daveski]

I've got some graphs of a six-axis IMU on my Bioloid MicroRaptor:

http://www.bioloid.info/tiki/tiki-index.php?page=MicroRaptor+Screenshots

You can see the accelerometers react very well to falling over. Whether or not they can sense it fast enough to prevent the fall is another question, one which I hope to answer in a few weeks...

- Jon

Wow, nice work Jon,
I think the accelerometer will sense movement fast enough, its just a case if the rest will follow and react fast enough.

The only experiance I have had with gyros is in helicopters and to make the most from the gyro reponse time the links from servo to control surfaces were kept short and stiff with a super fast corless digital servo that works at 0.07 sec over 60° working with 4.8v
Slop in the system wasn't an option. Thank the lord for solid state.

Love your micro Raptor and now I have a new link in my faveorites bookmarked..... http://www.bioloid.info/tiki/tiki-index.php

Keep up the good work Jon

Daveski

[daveski]

Ok, I think I found all the vids showing the feedback from the current load on the ankle servos.

humanoids balancing act
http://www.youtube.com/watch?v=HuU_GX6k8Gk

Bioloid on the slope (I think Bioloid was some how attached to the slope)
http://www.youtube.com/watch?v=HpFGROm-LQM

Don't push Bioloid
http://www.youtube.com:80/watch?v=EC5wZRckBQc

[smoothchat]

Wow, my videos made it onto You-Tube! I must be famous .

I had a lot of fun experimenting using the load sensors when making these videos.

Yes, for the sloping board video, the feet were prevented from sliding off the board by a couple of nails, but the physics and logic to find a balance were not affected. I could have used "rubber" under the feet to achieve the same result. So I don't feel that I was cheating in any way.

Re: Gyros, I'm not sure whether we can get a quick enough response to gyro measurements using the behaviour software. Even low level code might still have too much latency and result in a lot of hunting. It also doesn't help that the behaviour code doesn't support floating point operations. (Perhaps a future revision ??)

A gyro might need to be integrated into the CM5 directly IMHO. I hope I'm wrong

Good luck with your experiments.

[limor]

Wow, my videos made it onto You-Tube! I must be famous .

Yes.. Very impressive work smoothchat!! It is amazing that you managed to get it to counter-ballance to work even though the servo current measurement is very noisy and is only updated at the servo at a rate of 7hz AND you used the behaviour-control interpreter to do the control.

I had a lot of fun experimenting using the load sensors when making these videos. Yes, for the sloping board video, the feet were prevented from sliding off the board by a couple of nails, but the physics and logic to find a balance were not affected. I could have used "rubber" under the feet to achieve the same result. So I don't feel that I was cheating in any way.

You didn't have to tell us the "behind the scenes" details.

Re: Gyros, I'm not sure whether we can get a quick enough response to gyro measurements using the behaviour software. Even low level code might still have too much latency and result in a lot of hunting. It also doesn't help that the behaviour code doesn't support floating point operations. (Perhaps a future revision ??)

It is quick enough if written in native AVR. We were able to get the response from the Pepper board linked to a gyro at over 100hz and currently working on doing similar corrective action to get a bioloid to do a hand stand.

A gyro might need to be integrated into the CM5 directly IMHO. I hope I'm wrong

A gyro can be anywhere on the robot as long as it is in the right orientation and in the relevant segment of the many DOF of the robot.

[Kess]

It is amazing that you managed to get it to counter-balance to work even though the servo current measurement is very noisy and is only updated at the servo at a rate of 7hz AND you used the behaviour-control interpreter to do the control.

Blimey, were those balance videos really achieved using the standard Behaviour Control Programmer?! I've just started to try reprogramming my Bioloid (in C) to do basic balancing but haven't had much luck yet getting useable readings out of the AX-12 load sensors - I gather the load readings are not only slow and noisy but are also disrupted while the AX-12 is moving.

Are there any useful hints or advice anywhere on using the AX-12's load sensors for balancing? Even more useful would be to see one of the BCP files used to achieve those videos - are any of them in the public domain?

Kess

[MOHIT JINDAL]

I have the Bioloid premium kit that has 2 axis Gyro. Is the 6 axis gyro better than 2 axis ? Do I have to make changes in the motion file to make the gyro work or the Robot automatically detect it and balance itself ?