How to put the motion data and implement dynamically

[fairytalef00]

Hello,

In my simulation, I want to input motion data into a 6-axis manipulator and make it behave the same as the motion data.
I also want to know the force or torque acting on the end effector (or point on the manipulator where the force sensor is attached) when it moves according to the motion data.
The link length, mass, inertia, etc. of the manipulator are well-known.


In the real world, when the manipulator is moving, I have measured the angle of each of the 6 revolute joints every 0.005s at each of the six joints, which we can call motion(angle) data.
I want to substitute this into the exact same manipulator in CoppeliaSim,
so that it will have the same motion as the manipulator I experimented with in the real world.


However, my question is that measuring the angle at each time is strictly a position value, not a velocity or acceleration.
If I put the joint in dynamic mode and specify it as position control,
it will move according to the angle in the motion data at a given time, but the velocity at that position (or while going) will not be the same as the velocity seen in the real world.

Of course, we know it won't be exactly the same, but we didn't expect a completely different velocity for each position like we did with position control.
My understanding of position control is that it depends on max_torque or max_velocity, but is controlled with the goal of just getting to that position once at a given time.

In other words, if the motion data is given as '30 degrees in 0.005s, 60 degrees in 0.01s', then the position control might be 0 for the departure velocity and 0 for the arrival velocity, so it is different from the motion according to the motion data that I expected.



That was a long story, but the point is this.
In the real world, we store the robot's motion as 6-joint data over time,
is it possible to implement this in simulation in the same dynamic mode?

Actually, I already tried to implement the motion data in kinematic mode,
but as you know, it didn't make sense to me because I don't know the force or torque.

Thank you.

[fferri]

If you know the whole joint trajectory in advance you can pre-compute perfect velocity acceleration etc.

The problem of estimating velocity is so only if you want to do it in realtime (i.e. without knowing future values) then you can only estimate left derivative which is always "late" with respect to current trajectory point.

[fairytalef00]

Thank you!!

This is very helpful to me