Making it walk

[Anton Fosselius]

Is there anyone working on walking? this thread will focus on research and planning the construction of inmoov compatible legs. lets focus on energy efficiency, simplicity and price.

I think a good start would be a charli ii ripoff.

Some good reading:

http://www.ros.org/wiki/
http://www.openrtp.jp/openhrp3/en/index.html
http://www.romela.org/main/CHARLI:_Cognitive_Humanoid_Autonomous_Robot_with_Learning_Intelligence

[Amby Ambroise]

it s very difficult a big robot walking but i hope inmoov can walk

[gael langevin]

I'm already working on the legs, I have to say, I'm going to do it with the same approach as the arms. I don't really want copy other concept even though I would gain time, but as you maybe know, I do this for the fun and I like to come up with my own stuff. What I will do, is try to make something that can reproduce as much human movements, then hopefully we can use "something" that will control the different servos to balance both sides. To tell you how not conventional I'm working, I started with designing the ankle. Isn't that odd... Beside that,  I was wondering if I could install somekind of a small sensor close by the feet  to detect the shape of the floor so that the foot adapts to the right position for touching the ground during walking.
My main concern is the balance and coordination of all that though.
I will let you guys think of that since it isn't my best point, Grog on MRL might already have is own idea on that.

[Anton Fosselius]

Making it be able to balance requires alot of simulation during design. If you make the legs without simulations it will probably be impossible to make it walk. But on the other hand,if you design the legs first, we can then simulate it and tell you what needs to change in order to make it walk. Even when we got it balancing in the simulation, it might be possible to make it balance in the real world. There is no such thing as a perfect simulation. Search for "robocup humanoid" look at adult, teen and child classes. Some teams of ~20 persons have been working for years, and thats the best they can achive. (They can barely walk without falling).
If we make it walk, it will fall, and parts will break every time it falls. I think we will need alu rods as skeleton for the legs, there will be alot of force/torqe applied on them.

[James Bruton]

Hello, this is James from XRobots.co.uk . I've been following InMoov for a while.

I've recently built a bipedal android which had a simple method of control, It's about half-human size, and it uses R/C helicopter gyros for balance. It just about works ok although it's gait is not very human:

http://xrobots.co.uk/android-11/

http://www.youtube.com/watch?v=gYUHsg8LnUc

As you can tell by the name this was my 11th attempt, another one that worked quite well was Android 7, which used only three windscreen wiper motors to walk, and had no dynamic stability at all, this android was almost human sized at around 5 feet tall: http://www.youtube.com/watch?v=M6HZte7AF7k

My next plan is to move onto 'Android 12' which will be 3D printed, have more human-like leg joints, and more active stability using yet more gyros. The plan is to build the knee and ankles like some of the newer prosthetic limbs which react based on movement of the body and the rest of the leg, but are not linked to the wearers brain. This way I need only control the hips from a micro controller and the rest of the leg should react dynamically... it's experimental obviously.

[Anton Fosselius]

Thats some impressively small feet you got there! congrats on your good progress! My first humanoid had feets as big as half the height of the robot, it still had balancing issues ;)
Hope that your Android 12 turns ut good :D

[Anton Fosselius]

found the robot i was looking for before.

its known as the "The Cornell biped" its mostly mechanical magic and physics in action.

http://www.youtube.com/watch?v=P4aHzidIZ6M

and later built this:

http://www.youtube.com/watch?v=LDljbN7eJ5Y

[Christian Van Horn]

James, your robots are wonderful! I love what you have done on less than a million dollar budget, and I think we have a lot to learn from what you have accomplished. May I suggest incorporating InMoov into your design? With hands, arms, shoulders and more already up and working, it could save serious time in your next design. Grog over at MyRobotLab ( http://myrobotlab.org/ ), who has created InMoov control software, including vision, would love to help with the control software I'm sure, and we would all benefit! I'd love to know your thoughts, Chris

[James Bruton]

Hi Chris,

It's possible that could take place at a later stage, for now I need to invent a clever way to make it walk, it may not work at all in which case we'll be on to Android 13.

[Christian Van Horn]

I really liked you use of basic RC gyro technology. Sweet, and by the way, your costume work like the Iron Man suit is phenomenal! What exactly do you do for a living?

[Amby Ambroise]

hi james i have been following all your bipede robot

i think it s nice projet if you build a bipede humanoid likehuman

so it is possible you put open source your work like gael?

thank you

[James Bruton]

Yes for the 3D printed parts, and once it works, I will publish everything. The previous attempts were hand made so there we no plans or CAD files.

[gael langevin]

Hi James, Welcome to InMoov,
Wow, you have done an amazing job with that 5 feet tall robot. the mechanism is very impressive. Your knowledge in balancing the robot would be a great source of info, if you are willing to share it with us. The concept of building a robot leg without motors and only dynamicaly is very interesting because the movements would be much realistic. The video posted by Anton shows it very well. It also could drastically simplify the construction. Unfortunatly I have started working on the ankle and foot mechanism and haven't gone in that direction...
That's the problem and the good thing with seing others projects.
I hope you will help us with programming.

[James Bruton]

Thanks Gael, I will share anything I can, I generally take step by step photos of things as I build them, you can see all of the Android in detail on this page: http://xrobots.co.uk/androids/

Most are hand made however so there are no further plans, but you can see roughly what the concept is and how things fit together.

[gael langevin]

@James,
thanks for the link, I will have a good time studying it. Is the program done for an arduino microcontroller? I might submit it to Grog on MRL if it's okay with you.

[James Bruton]

No, I used Picaxe for everything so far. The intention being to show that control is very simple and that's all you need:

www.picaxe.com

[James Bruton]

Here's a really good video about some robotic/prosthetic ankles. Check out the guy wearing one on each leg about half way through: 

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

[James Bruton]

Sorry, that link should be: http://www.youtube.com/watch?v=PbfcDlZHtjY

[gael langevin]

It is totally amazing! The ankle is sensing the force input to respond in accordance.  Another direction again. Unfortunatly I don't think the ankle I'm building now will be as reactif.
Thanks for the video, James.

[B Stott]

Hi James,

I just read through this thread and my brain over loaded on your xrobots.co.uk/androids link. You have really put some quality effort into that. I'll be studying it once I get some senses back. :-)

Good Job.

[Haydn Davies]

Hi James,

Quite a coincidence.

I was looking at your site only yesterday for some info about Jesmonite. I‘ll be posting on your forum with some questions later.

I remember Picaxe !!! Must be 5 to 10 years since I got into them. They seemed to be the easiest microcontroller to program in those days and a good price. (I was brought up on MS Basic, and found similarities.)

But I don’t have a clue as to WHY I needed a microcontroller. I seem to have no trace of the project in the attic or in my mind…

 Haydn

[Alan Hart R & D Robotics]

What sort of weight does inmoov need to have in the leg say from the knee down that is currently on the ground. Would two to three kg be enough to keep it stable.

 

Alan

 

 

[gael langevin]

Don't know if that weight really matters. I think the equilibrium controlled by chip is essential and doesn't have to be done by over charging the legs.

2013/6/25 Alan Hart R & D Robotics <ontracke...@hotmail.com>

What sort of weight does inmoov need to have in the leg say from the knee down that is currently on the ground. Would two to three kg be enough to keep it stable.

 

Alan

 

 

--
You received this message because you are subscribed to the Google Groups "InMoov" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inmoov+un...@googlegroups.com.
To post to this group, send email to inm...@googlegroups.com.
Visit this group at http://groups.google.com/group/inmoov.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Alan Hart R & D Robotics]

Correct me if I am wrong, but as we shift our weight onto the leg that is standing our weight on our other leg is reduced.  the walking leg is then raised moved forward backwards etc and the placed back down.  The weight is then shifted to this leg and the other leg does the same thing.  Using a pump and liquid this weight can be transferred back and forth for walking or equalised in both legs when standing still.  Thinking along the linmes of ballast and centre of gravity conditions in flight.

Of course the chip could control all this.

 

 

 

Alan

 

 

 

 

[James Bruton]

I think you'll have a job to pump the water backwards and forwards at the right rate / fast enough for a normal walking motion, and dynamically distribute the mass so that momentum doesn't build up causing it to overbalance. You'd be better off measuring the direction and velocity of travel as the android tips and dynamically tuning the motion to it's natural oscillation based on the coefficient of the speed of its actuators and it's overall size & the effect gravity has on it - basically like the way humans actually walk using the inner ear to measure motion and stay balanced.

Here's my own attempt so far, my next steps are to slow down the oscillation using the speed and acceleration settings of the servo controllers, and then try to get some small steps in: http://www.youtube.com/watch?v=jDOqpwcKhhg

More pictures and construction details at: http://xrobots.co.uk/android12/

Sorry to post an unrelated project here, obviously the principles could be used in other systems (there are other ways to do it too using an accelerometer/gyro pair mixed via a low/high pass filter respectively).

[gael langevin]

Very interesting approach. I have never thought of that for walking.
I don't think our leg is getting lighter as we pull it up for to walk, is it really? Maybe you are right, i won't prove the contrary without knowledge.
Thinking outloud: the whole body shift to counterbalance the weight of the lifted leg, and thats it.  I can't imagine such a transfer of weight from one leg to the other, specially when we run. Or even when a leopard runs.

[Alan Hart R & D Robotics]

http://i00.i.aliimg.com/wsphoto/v0/572991848/Free-Shipping-New-Portable-10L-Camping-Handle-Collapsible-Folding-font-b-Water-b-font-Bucket-font.jpg_250x250.jpg

 

 

On Monday, July 1, 2013 12:47:57 PM UTC+10, Alan Hart R & D Robotics wrote:

On the shifting of weight from one leg to another

Think of something like a collapsable water container used in 4WDing in each leg. As one leg is raised the water is forced by a small piston into the other container which then expands and back and forth.

 

 

[Alan Hart R & D Robotics]

Probaly weight may not be the correct wordingm should have been more centre of gravity or balance.

 

Quick overview is in this animation pdf

 

http://www.algarcia.org/AnimationPhysics/BalanceTutorial.pdf

[gael langevin]

Thanks Alan.

We can see in these animations that weight doesn't shift from one leg to the other but it is all through counterbalance with gravity. I think it should be workable with software if the mechanism can respond quickly enough to the demand during the different dynamic moves.

2013/7/1 Alan Hart R & D Robotics <ontracke...@hotmail.com>

Probaly weight may not be the correct wordingm should have been more centre of gravity or balance.

 

Quick overview is in this animation pdf

 

http://www.algarcia.org/AnimationPhysics/BalanceTutorial.pdf

--

[gham...@unitec.ac.nz]

Hello Rossano
Mine is going on a Segway (the old one with 4 wheels) with VR glasses (Vuzix Wrap 1200VR) for robot head movement.

On Tuesday, 5 March 2013 08:43:48 UTC+13, Anton Fosselius wrote:

Is there anyone working on walking? this thread will focus on research and planning the construction of inmoov compatible legs. lets focus on energy efficiency, simplicity and price.

I think a good start would be a charli ii ripoff.

Some good reading:

http://www.ros.org/wiki/
http://www.openrtp.jp/openhrp3/en/index.html
http://www.romela.org/main/CHARLI:_Cognitive_Humanoid_Autonomous_Robot_with_Learning_Intelligence