Information flow from ROS to real UR Robot #489
Replies: 1 comment 2 replies
-
To understand what is happening, you'll first have to understand how ROS's joint trajectory controller (JTC) works. I\ll try to summarize that very briefly. When a trajectory is being sent to the controller, the controller builds a spline interpolation based on the trajectory's setpoints. This way, there is a continuous representation of each robot joint w.r.t. joint position, velocity and acceleration. The ROS control loop takes a sample point at the current time from this spline curve which is the current command. Based on the robot's interfaces, this current command will be sent to the robot. This happens in every control loop, so in the case of a UR e-series robot every two milliseconds. As you can see from the controller configuration this driver uses positions being commanded to the robot. Those positions are then being sent to the robot using the Everything said here could also be implemented using velocity commands, in which case |
Beta Was this translation helpful? Give feedback.
-
Can anyone please help me understand, how the trajectory waypoints are send from ROS to the UR controller?
What I mean is: what all information is being transmitted -- waypoints, moveJ, moveL, TCP, APO, Speed, Acceleration, Override etc. ?
All the trajectory waypoints that MoveIt has planned are being send to the robot controller at once or a new waypoint is sent once the old waypoint is reached?
Beta Was this translation helpful? Give feedback.
All reactions