| Comments on CHRONO example robot_delta_learn.r3d |
| Actuators which learn motion. |
| What to do: load the file, open the control panel, press the play button on the animation timeline (watch at least 6 seconds of simulation!). Be sure that time slider is at frame 0 (beginning) when you load the project. |
Here you can see the simulation of a 3 degrees-of-freedom (3DOF)
parallel robot, a novel 'Delta(tm)' manipulator which is used mainly
for fast pick-and-place operations.
Note that the robot is in 'learn' mode, that is, its motion
does not happen because the 3 linear actuators are moving it, but rather because
its end effector (ex.the gripper) is 'pulled' by a reference which
moves in space, hence obtaining an inverse-kinematic movement.
In fact, if you open the property window of the 3 linear actuators ('linear actuator'
objects in 'System' level), you see that their 'learn mode' flag is ON,
meaning that they are switched off and they just record their rotation
in the 'd(t) extension' ChFunction.
By the way, since motors are in 'learn motion' mode, how is this device moved?
Note that there is a 'TEACHER ..' link in System level, which constrains
the position of the tool tip to a moving marker. Hence the moving marker,
performing as a 'teacher' which pulls the robot arm, is in 'ground', and
its motion has been created with the default keyframing tools of Realsoft (you can
see and edit its choreographies).
Interesting remark for engineers: after you performed 6 seconds of this simulation,
the motors have learnt their rotation, so you could go back to the beginning
of the simulation and do the following: set 'disabled' ON for the 'TEACHER ..' link,
set 'learn mode' OFF for the three 'linear actuator' objects. Than play simulation again,
and you see that robot motion is the same (but this time is imposed by telescopic
linear actuators, so you can use Probes to record the needed forces and speeds).
© DeltaKnowledge 2007