I came up with this project because I wanted to explore the intricacies in highly dynamic human robot interactions. Robots now can carry things themselves, so I wanted to investigate how can they collaborate with humans on these tasks with vision-free sensing (force-torque) 

The goal is to teach a robot with a rigid end-effector to collaborative "pinch" objects with a human and move it in space. We have achieved success in 1d translations for various objects(both push and moving away). Our goal is to generalize it to move in 3d. 

​

github