Real-Time Physics Simulation Forum

Re: Bullet vs ODE for simulating robotic systems

2014-09-04T05:40:10+00:00

Basroil, you are reading the code of the latest https://github.com/bulletphysics/bullet3 trunk/master version, right?
I recall hooking up the MLCP joint feedback, but haven't looked into it recently.

Yes, github bullet3 master with commit af5883c being the last modification to that file. I would have assumed that joint feedback would have been done within a solver finish function but none exists. Is the underlying SI code (inherited) being used in place of a dedicated solver finish and i'm just confusing the need for a custom version? (very likely, since it it's mostly cleanup)

Statistics: Posted by Basroil — Thu Sep 04, 2014 5:40 am

Re: Bullet vs ODE for simulating robotic systems

2014-09-04T03:54:29+00:00

You can use setJointFeedback on each constraint, similar to ODE, see
https://github.com/bulletphysics/bullet ... aint.h#274
Rereading one more time and following through all affected code, I noticed that only sequential impulse and multibody solvers seem to have joint feedback actually implemented (modifies 4 variables from m_appliedForceBodyA to m_appliedTorqueBodyB in some way, looks like an exact copy for all intents and purposes). Would it be safe to assume that the c.m_appliedImpulse for MLCP can be used identically to the one found at the end of group solve finish in multibody and sequential impulse?

Basroil, you are reading the code of the latest https://github.com/bulletphysics/bullet3 trunk/master version, right?
I recall hooking up the MLCP joint feedback, but haven't looked into it recently.

Statistics: Posted by Erwin Coumans — Thu Sep 04, 2014 3:54 am

Re: Bullet vs ODE for simulating robotic systems

2014-09-04T02:53:01+00:00

You can use setJointFeedback on each constraint, similar to ODE, see
https://github.com/bulletphysics/bullet ... aint.h#274

Rereading one more time and following through all affected code, I noticed that only sequential impulse and multibody solvers seem to have joint feedback actually implemented (modifies 4 variables from m_appliedForceBodyA to m_appliedTorqueBodyB in some way, looks like an exact copy for all intents and purposes). Would it be safe to assume that the c.m_appliedImpulse for MLCP can be used identically to the one found at the end of group solve finish in multibody and sequential impulse?

Statistics: Posted by Basroil — Thu Sep 04, 2014 2:53 am

Re: Bullet vs ODE for simulating robotic systems

2014-09-03T04:24:25+00:00

You can use setJointFeedback on each constraint, similar to ODE, see
https://github.com/bulletphysics/bullet ... aint.h#274

Misread his statement as saying "by a specific body" rather than "on a specific body"

Edualbu, disregard my other answer, it was just my mistake.

Statistics: Posted by Basroil — Wed Sep 03, 2014 4:24 am

Re: Bullet vs ODE for simulating robotic systems

2014-08-23T19:53:06+00:00

1. (Can Bullet Engine) Find out how much force is applied on a specific body and what is the direction (vector) of that force, in order to implement "force" or "pressure" sensors. Something equivalent to ODE's dJointFeedback struct.
Does it? Not really in 2.82. Can it be modified to include it? Sure, though the difficulty will vary depending on what you want to do. Applied forces are based on the body center, so if you need the force vectors at the contact point (rather than equivalent forces), you'll have to work backwards a bit (might impact performance). You could also intercept the constraints results themselves, though you risk problems if you aren't careful.

You can use setJointFeedback on each constraint, similar to ODE, see
https://github.com/bulletphysics/bullet ... aint.h#274

Statistics: Posted by Erwin Coumans — Sat Aug 23, 2014 7:53 pm

Re: Bullet vs ODE for simulating robotic systems

2014-08-08T04:13:59+00:00

1. (Can Bullet Engine) Find out how much force is applied on a specific body and what is the direction (vector) of that force, in order to implement "force" or "pressure" sensors. Something equivalent to ODE's dJointFeedback struct.

Does it? Not really in 2.82. Can it be modified to include it? Sure, though the difficulty will vary depending on what you want to do. Applied forces are based on the body center, so if you need the force vectors at the contact point (rather than equivalent forces), you'll have to work backwards a bit (might impact performance). You could also intercept the constraints results themselves, though you risk problems if you aren't careful.

2. Find out the position and normal vectors of contact points/edges of collision detection. This is necessary to provide visual feedback about the collision to our users. Something equivalent to ODE's dContactGeom.

There are plenty of contact demos included in the bullet code base, and there's just about every case covered.

3. Find out how much torque/force is currently used by a hinge/slider motor while moving or maintaining the current position (motor force feedback).
..."

These points are very important for my project. Do someone has any information about that?

Thanks for the help,

Eduardo

Since motors are just constraints, the question is identical to question 1. Though, you might not even have to worry about that if you use imperfect motors (non-infinite torque), since any proportional motor can give you basic feedback simply in the motor error (low accuracy, but usually good enough).

Statistics: Posted by Basroil — Fri Aug 08, 2014 4:13 am

Re: Bullet vs ODE for simulating robotic systems

2014-08-06T18:56:17+00:00

Thanks a lot, Basroil.

I'm also interested in the questions Yvan Bourquin asked about Bullet a long time ago. Here are the questions:

"...
1. (Can Bullet Engine) Find out how much force is applied on a specific body and what is the direction (vector) of that force, in order to implement "force" or "pressure" sensors. Something equivalent to ODE's dJointFeedback struct.

2. Find out the position and normal vectors of contact points/edges of collision detection. This is necessary to provide visual feedback about the collision to our users. Something equivalent to ODE's dContactGeom.

3. Find out how much torque/force is currently used by a hinge/slider motor while moving or maintaining the current position (motor force feedback).
..."

These points are very important for my project. Do someone has any information about that?

Thanks for the help,

Eduardo

Statistics: Posted by edualbu — Wed Aug 06, 2014 6:56 pm

Re: Bullet vs ODE for simulating robotic systems

2014-08-06T01:30:22+00:00

I'd like to know if the Dantzig Solver is already implemented. I'm also working if robotic simulation and I'm trying to choose between ODE and Bullet.

It has been for a while, and for low DoF systems (<20 joints or so) it works faster than PGS with the high iterations need for powered joints. ODE is no longer the go-to choice for robots. I've personally modeled robots using AX-12 motors (including motor load characteristics) using just Bullet hinges and Dantzig. The only issue right now seems to be that (according to the comments) friction doesn't work the same way as in PGS. You can also check to see if featherstone works for you, though it's a bit harder to set up real motors using it.

Statistics: Posted by Basroil — Wed Aug 06, 2014 1:30 am

Re: Bullet vs ODE for simulating robotic systems

2014-08-06T19:13:18+00:00

Hi,

I'd like to know if the Dantzig Solver is already implemented. I'm also working with robotic simulation and I'm trying to choose between ODE and Bullet.

Best regards,

Eduardo Albuquerque

Statistics: Posted by edualbu — Tue Aug 05, 2014 11:11 pm

Re: Bullet vs ODE for simulating robotic systems

2010-06-30T14:37:45+00:00

Putting in another vote for improvements for robotics. Another student in my research lab has discussed these same issues about mass ratios and such. My own issues are with working at small scale, so I use units as cm instead of meters, but I keep gravity relatively weak since it seems the constraint forces can't keep up with the fully scaled gravity.

Anyway, if the Dantzig will solve at least some problems, I vote for putting this when at the soonest opportunity.

Meanwhile, thanks much for the hard work on Bullet.

Statistics: Posted by tompalmer — Wed Jun 30, 2010 2:37 pm