How is Bioloid Biped's Stability W/O Gyros / Accelerometers?

[NullARC]

I was wondering if anyone could tell me how well a Bioloid Biped is able to balance without gyro or accelerometer help?

Does anyone know how these sensors might be integrated into a Bioloid kit to enhance stability and balance?

Thanks for your input!

- Marc

[Ray]

that's why I am not willing to invest to the bioloid even though it has many attractive features, e.g. high torque, C programming, ir sensor, ......

but as some of the hobby said, it is a lego, not just humanoid,..... you are not in the right place.....

so, I give it up! and look for better robot then...

[inaki]

There are two approaches to insert a gyro in an humanoid. One is a gyro inserted directly in the servo connection. In this case the gyro automatically adjust PWM pulses to compensate servo position. This is used for example in KHR1 (and in most RC vehicles that use gyros). This method does not require any management from the controller. It cannot be used on Bioloid because it uses a data bus and there is no visible PWM signals to manage. The inconvenient of this approach is that it uses one gyro for one servo. You cannot use one gyro for two or more servos and of course you cannot use the gyro to take complex decisions.

The other approach is read the gyro(s) from the controller and adjust servo(s) positions accordingly. This approach would work on Bioloid. Obviously it would require a gyro connected to Bioloid bus with a suitable hardware.
I am positive that kind of sensors will be available for Bioloid robots in the near future because there are already many people working with Bioloid bus with success.

Without gyros Bioloid can walk pretty well and you can even perform some *stability* control using 'servo load' parameter (I think this has been tried with success by at least one Robosavvy forum member).
Althoug using gyros would be much better solution, there is no gyro sensors currently available for Bioloid.

[HarryBotter]

I agree that balance is a necessary feature of any humanoid whether using a gyro, accelerometer or both. I've done some fooling around with balance using both a gyro and accelerometer on a wheeled bot and am really anxious to try that on the Bioloid once the communication issues are solved.

I've posted a short video on the Robot magazine forum that shows my Bioloid humanoid balancing.
http://www.botmag.com/forum/showthread.php?t=313
The balancing itself is a bit slow but I'm going to see if I can use it to improve walking on a carpet.

Harry

[pepperm]

Harry

The video looks great. Is there any chance we could have a look at your code please to see how it is done?

Mark

[Pev]

Well I hope this helps. This is a video of my Bioloid walking, no gyros no stabalisation software just the bog standard stuff.

http://robosavvy.com/Builders/Pev/18DOF.wmv

Now this may be controvertial but I actually think he is more stable than my RN-1 with a Gyro!!!!!!!!! okay its only my opinion and I am very happy with the RN-1s performance too. I think the extra 2 DOF that rotate the legs make Bioloid inheritantly more stable but again thats just my opinion.

Hope this helps

Pev

[NullARC]

Pev,

Great work!! I agree, I think it's stability is incredible! The speed at which it is moving is pretty impressive. To move that quickly with that stability is great. That's another check mark in the "Pro's" column in my Bioliod, AT-01, KHR-1HV quandry. The more I study and research the more I lean towards the Bioloid system.

Is that a custom gait that you designed with the motion editor, or was it written directly in C?

[HarryBotter]

Pepperm, I've got several iterations of my program and can't easily figure out whiuch is the right one. When I get a moment I'll figure it out. Why don't you try it yourself, it's not really that hard. When you start experiencing load at the ankle joint just move in the opposite direction.

Pev, you're absolutely right, these bots are very stable without any assistance right out of the box. I sure wonder how much more could be achieved with gyros/accelerometers.

Harry

[Pev]

Hi,

The gait is actually a Robotis supplied gait on the last update CD, which I did a few minor tweaks to. I also wonder where any additional stability could be gained with gyros etc but of course the challenge would be so keep the stability in sloped or uneven environments. Now some feedback would be a must in those environments

Pev

[Ray]

stability issue is dominant especially the robot is walking in a step.
In the video, the foot is moving so that the toe is near the floor and thus the stability is better.
If the foot is lift upper, the stability is getting bad.

[limor]

HarryBotter,
the current-sensing balancing act is very cool.
Please upload the video to youtube or upload it here because to watch that video requires registration.
In the video it seems that your corrective ankle servo updates are at about 1-2 times per second. Have you tried to do the update at a higher rate ?

[HarryBotter]

Thanks limor, it's quite neat to be able to do that without gyros/accelerometers.

I'd rather not post the video anywhere else because Smoothchat actually did this before me and I don't want to steal his thunder.

I'm making the updates as fast as possible and thought the slow speed was probably due to the interpreter. If I can figure out which program I was using I'll post it and you can take a look.