Robotics env how to get position of object and gripper


#1

In the paper multi-goal reinforcement learning challenging robotics environments and request for research

(https://arxiv.org/pdf/1802.09464.pdf)
they are using grip position and object position. They have not explained in the paper though how they obtain this information. Is there any algorithm or code that we can get the position of the object/gripper from a installed camera.


#2

I think you can do some thing like : https://www.pyimagesearch.com/2015/01/19/find-distance-camera-objectmarker-using-python-opencv/

I think you should find out the distance of object and move the grip along synchronously.
More over then you have to process gripper and make a decision to make it grip the object.

Thanks,
Rahul


#3

Josh Tobin:

https://blog.openai.com/tag/josh-tobin/

He is having a get together at a Texas University, USA. April 27 2018:

https://apps.cs.utexas.edu/calendar/events/278722

AGi uses the contraction and expantion of object when moving to them or away.