how Robot keeps balance??

Korean company maker of Robot kits and servos designed for of articulated robots. Re-incarnation of Megarobotics.
6 postsPage 1 of 1
6 postsPage 1 of 1

how Robot keeps balance??

Post by elikpr » Sat Jul 02, 2011 3:53 pm

Post by elikpr
Sat Jul 02, 2011 3:53 pm

Hi guys

I want to ask about how robot keeps balance while walking(motion)? Is there any kind of theorem or something according to which whatever robot fulfills it does not fall? Or.....

Thanks in advance
Hi guys

I want to ask about how robot keeps balance while walking(motion)? Is there any kind of theorem or something according to which whatever robot fulfills it does not fall? Or.....

Thanks in advance
elikpr offline
Robot Builder
Robot Builder
Posts: 14
Joined: Tue Apr 19, 2011 2:30 pm

Post by limor » Sat Jul 02, 2011 10:39 pm

Post by limor
Sat Jul 02, 2011 10:39 pm

thats a big question.
The kit humanoids don't fall because they have big feet and most of the gaits you see on youtube are designed by trial and error to ensure the robot doesn't fall if the floor is consistent.
on uneven terrain gyros are used with hobby humanoids to compensate for small aberrations on the floor.
real dynamic walking is complicated. most of the human size humanoids you see on youtube rely on ZMP principles.
http://en.wikipedia.org/wiki/Zero_Moment_Point
thats a big question.
The kit humanoids don't fall because they have big feet and most of the gaits you see on youtube are designed by trial and error to ensure the robot doesn't fall if the floor is consistent.
on uneven terrain gyros are used with hobby humanoids to compensate for small aberrations on the floor.
real dynamic walking is complicated. most of the human size humanoids you see on youtube rely on ZMP principles.
http://en.wikipedia.org/wiki/Zero_Moment_Point
limor offline
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 1845
Joined: Mon Oct 11, 2004 1:00 am
Location: London, UK

Post by i-Bot » Sun Jul 03, 2011 12:27 pm

Post by i-Bot
Sun Jul 03, 2011 12:27 pm

As Limor describes, there are two models of stability.

In the static state, and also assumed for the "stationary walk", the centre of gravity must lie within the supporting polygon. The supporting polygon is the foot or feet if two are on the ground. You can use a marble under a perspex sheet and move it around to find where the centre of gravity is. The degree of stability is determined by how high the center of gravity must be lifted to move it outside the support polygon. Thus the further the centre of gravity is from the edge of the polygon, and the lower the COG the better for stability. However too stable and you take a lot of energy to move and move quite slowly.

The problem with the "stationary walk" is that it makes no account of inertia while moving. The Zero Motion Point is created to allow for that. Now the Zero Motion Point must be within the supporting polygon, and the centre of gravity, may or may not be.

There are still external disturbances which can change stability, so some external feedback is usually also applied with Gyro, IMU and force sensors.
As Limor describes, there are two models of stability.

In the static state, and also assumed for the "stationary walk", the centre of gravity must lie within the supporting polygon. The supporting polygon is the foot or feet if two are on the ground. You can use a marble under a perspex sheet and move it around to find where the centre of gravity is. The degree of stability is determined by how high the center of gravity must be lifted to move it outside the support polygon. Thus the further the centre of gravity is from the edge of the polygon, and the lower the COG the better for stability. However too stable and you take a lot of energy to move and move quite slowly.

The problem with the "stationary walk" is that it makes no account of inertia while moving. The Zero Motion Point is created to allow for that. Now the Zero Motion Point must be within the supporting polygon, and the centre of gravity, may or may not be.

There are still external disturbances which can change stability, so some external feedback is usually also applied with Gyro, IMU and force sensors.
i-Bot offline
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 1142
Joined: Wed May 17, 2006 1:00 am

Post by elikpr » Mon Jul 04, 2011 6:02 am

Post by elikpr
Mon Jul 04, 2011 6:02 am

Can you give an example code(simple walking, turning and so on )so that I could understand more clearly?

Thank you
Can you give an example code(simple walking, turning and so on )so that I could understand more clearly?

Thank you
elikpr offline
Robot Builder
Robot Builder
Posts: 14
Joined: Tue Apr 19, 2011 2:30 pm

Post by i-Bot » Mon Jul 04, 2011 11:01 am

Post by i-Bot
Mon Jul 04, 2011 11:01 am

The existing Motionbuilder files are the examples. They have been hand crafted to meet the stability criteria.

You will find some moves always keep the centre of gravity within the supporting polygon. These moves do not fall when you step through the moves.

Moves which work OK as a complete motion, but fall when stepped through are usually due to the reliance on meeting the requirement to keep the zero motion point inside the supporting polygon, but the centre of gravity is outside.
The existing Motionbuilder files are the examples. They have been hand crafted to meet the stability criteria.

You will find some moves always keep the centre of gravity within the supporting polygon. These moves do not fall when you step through the moves.

Moves which work OK as a complete motion, but fall when stepped through are usually due to the reliance on meeting the requirement to keep the zero motion point inside the supporting polygon, but the centre of gravity is outside.
i-Bot offline
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 1142
Joined: Wed May 17, 2006 1:00 am

Post by elikpr » Thu Jul 14, 2011 1:51 pm

Post by elikpr
Thu Jul 14, 2011 1:51 pm

I have read about ZMP, as far as I understand to use ZMP in my robot I need to find out Center of mass. But In my case My robot does not have any additional stuff like Gyro to implement ZMP. Is it possible without gyro or other stuff to implement Zero Moment Point algorithm?( Sorry if I am giving strange questions, I am really new in robotics.)

Thank you in advance
I have read about ZMP, as far as I understand to use ZMP in my robot I need to find out Center of mass. But In my case My robot does not have any additional stuff like Gyro to implement ZMP. Is it possible without gyro or other stuff to implement Zero Moment Point algorithm?( Sorry if I am giving strange questions, I am really new in robotics.)

Thank you in advance
elikpr offline
Robot Builder
Robot Builder
Posts: 14
Joined: Tue Apr 19, 2011 2:30 pm


6 postsPage 1 of 1
6 postsPage 1 of 1