Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5

Real-time teleoperation of a real robot (Mecademic)

#1
Hello everyone,

I trust you are all doing well.

I am currently working in a school project involving the retrieval of data from a Motion Tracker (SMARTTRACK from ART). The output comprises a 6x1 vector containing both the position and orientation of the body.

Successfully, I have developed a script capable of retrieving data from the tracker, generating a pose from the 6x1 vector, and simulating the acquired movements on a virtual robot using the "SetPose()" function.

Now, I am working on the next step, which involves real-time simulation of these movements on an actual robot (Mecademic R3). In my initial attempts, I created a target for each iteration of the loop and used MoveJ to maneuver the real robot to that target. However, this approach resulted in a discrete and non-smooth motion.

Subsequently, I experimented with a single target, updating its pose in each loop iteration. Unfortunately, this too resulted in the same issue.

Is there a proper method to teleoperate movements from the virtual robot to the real robot? essentially synchronizing their movements. Is it possible to use "setPose()" on the real robot? Can this issue be only related to the use of mecademic robot? 

I appreciate your assistance in advance.

Best regards,
#2
You should split the movements into small linear or joint movements (MoveL or MoveJ) and always send the pose that is close to the current location but moves towards the new position you want to move to. 

This allows you to keep providing new movements as you calculate the new position the robot should move to.
  




Users browsing this thread:
1 Guest(s)