bipedal robots

solo220

Hi guys i am new to this forum and wondering if any one is dealing with bipedal robots

like the ones used IN "ROBOCUP" competition in " http://www.robocup.org/ "
or like honda's AZIMO (in a smaller scale/size :-) ).

like all my unfinished home projects i was building one but faced some mechanical/stability issues
because i made it big (1 m tall) from the beginning .

until i realized that big/heavy ones have big problems regarding stability and needed servo motor torque
(burned two of them :-( ).

waiting for your replies "lebbotgeeks" .

Nabs

What exactly do you need help with, Solo220?

Ayman

I personally think that the costs of having a bipedal robot outweighs its benefits. For example a robot on wheels or tracks would be much cheaper to build and much more robust in terms of movement.

A bipedal robot would need much more work, money and sophistication to make it have balance while walking and standing too just like a human being. The only benefit I see in it is its ability to climb stairs and of course its look and feel as a robot.

But of course it is really cool and fun to create a humanoid with real human like legs. J4D is the robotics guy here, I think he would love to build a bipedal robot some day, I remember we had a discussion regarding this topic previously.

Btw, did you follow some tutorial online like those on instructables.com ? Did you come up with the design on your own?

solo220

@nabs well i am an electrical engineer and my weakness is in mechanical issues , i found that stability and balance is not based always on electrical sensors (gyro , accelerometers,.... ) but also on some body design and Center of Gravity , so i am looking for some idea's for a skeleton which is mechanically stable and could be home built .

@aymanfarhat a bipedal robot is the ultimate tool for any task that needs a human to do it , in a human world every thing is made to accommodate for human body ...
and beside i like challenging my self with new ideas and technology :-) , so i see robots on wheels as an old /kids toy unless you have something else to mount on them ;-) .

in my project i designed the legs only on autocad and then made them of plexi glass which proved to be very heavy compared to aluminium and fiber glass :-( .
then i used some (10) heavy large chinese servo motors with 13Kg/cm (as said in datasheet) but they needed about 5A/7.2V so i used a lipo battery , suddenly 2 of them started smoking :-( .

and by the way you can buy a futaba bipedal for 1000$ online price.

or a smaller one called "isobot " for 300$ (i did have one)

but my target is to build one with at least 1m long :-p .

Ayman

@aymanfarhat a bipedal robot is the ultimate tool for any task that needs a human to do it , in a human world every thing is made to accommodate for human body ...
and beside i like challenging my self with new ideas and technology :-) , so i see robots on wheels as an old /kids toy unless you have something else to mount on them ;-) .

Yes that is true but remember that robot is not really a robot by just being a machine that moves, software plays a huge role. In my opinion a machine with an MCU and on wheels with useful sensors and intelligent software written for it is much more of a robot than another machine with legs that just walks.

The way I see it, by the time and effort you spend on making the mechanics work for the bipedal robot, you could be building some really neat robot on tracks with some neat software.

Btw there exists some ready made mechanical humanoid legs you can buy and install into your robot without the need to re-invent the wheel.

But I am not sure about the size, maybe what you're aiming for is something as big as Asimo?

solo220

i agree with you 100% if my work is for some real project from which i should earn money .

but i think it's better for me to do the bipedal because of many reasons:

1-building that bipedal robot (1 m long) is like a hobby (maybe a fantasy) so i don't care for
feasibility or work+money/benefit ratio .

2- i'm a hardware guy so software and Artificial intelligence is not my type( despite some assembly,vb,java(android),c++,...).

3- i like mechatronics and automation (which what i do for living).

4- a bipedal will be a machine control challenge for me from which i could learn some new things in control loops.

and i have a bipedal that is better than BRAT (the bipedal that you gave me the link for) it is called "ISOBOT"
www.isobotrobot.com but thank you any way .

helixgecko

i salute your intentions and determination first.
building a humanoid robot from scratch is a big deal for a personal project with no budget and sponsors.
i would normally recommend to put effort on software; computer vision and interaction.
but you said you are interested in the mechanical part, which basiclly should be built with lego to test robustness and efficiency and then switch to autocad design and so on.

solo220

thank you helix , may be i will try to build a skeleton with Lego Mindstorm parts so i can understand skeleton key points.

my problem (as i see) that a walking bipedal skeleton (without actuators moving ) have to be almost in equilibrium in all its positions or it will fall ,

i mean try to stand on one leg you will fall if nothing balance the other leg , your body balance system is always trying to maintain the center of gravity around the pelvis, so this not always about muscles it is about equilibrium , this was my first error .

MY PROBLEM NOW IS HOW TO DETERMINE THE DIMENSION FOR THE PARTS BETWEEN THE JOINTS , AND THE WEIGHT DISTRIBUTION to maintain equilibrium while walking.

i thought that if someone is working on a similar project , he may point me to some skeleton design criteria or do and don't.

solo220

this is the autocad drawing that i made for one leg

parts are made of 2.5mm plexi glass , i cut them using a modified Rolland "CAMM2"
desktop CNC machine

helixgecko

i am sure there are alot of articles about equilibrium for humanoid, which can give you some plans to try.
The drawing shows you have at least 4 degrees of freedom for a leg; i recommend to start simple and then advance.

solo220

your right i think it's better for me to start building a smaller lighter one,

i recommend to start simple and then advance

may be i was a little in a hurry :-) , and i was giving my big servos too much trust :-p ,

thx anyway i apreciate your time and advices.

solo220

or i can cut the time and try to remove I-SOBOT brain and check how i can move it without he falls .

solo220

here it is the LOWER part of the leg ready and built

any comments are very welcomed.

helixgecko

I can't see how the leg bends for moving forward, the mechanism looks awkward :/

solo220

I am sorry it is my fault :-( there is two more servo's in the upper part not fixed in the picture i will try to post a more complete picture, may be with the two legs.

solo220

This is a 3dmax animation for the robot walking sequence as i imagined it

http://www.youtube.com/watch?v=6GtlEXsm2pE&feature=youtu.be

solo220

one complete leg controller test.

http://youtu.be/hj0Mu9WIRoE

mesa177

solo220 wroteparts are made of 2.5mm plexi glass , i cut them using a modified Rolland "CAMM2"
desktop CNC machine

I have a question (hope I'm not hijacking the thread or anything), do you own the CNC machine, or do you go to a specific shop? If you own the machine, do you mind telling me more about it (material limitations, size, cost, etc...)?

PS: Nice design :)

solo220

@mesa
Actually i bought the machine from some shop in "ard jalloul" (as they call it) for around 50 dollars and removed the broken electronics board than installed a pic microcontroller with other stepper motor peripherals
than using VB6 i made a program to control the motors (still need some tuning).
its is a ROLLAND CAMM-2 MODEL PNC2300a
http://www.zelkovaltd.com/Plotterworld/store/stock/Roland/Brochure/Camm-2.htm
2.5D , around 25X25 cm work area, i did some plexiglass and wood on it only.