Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Retargeting for Robots
Robots can already fulfill a lot of helpful tasks, like pick-and-place of rigid objects or parts assembly, which is nowadays the state-of-the-art in the industry. Nevertheless, they cannot execute more complex tasks with the same agility humans have (yet!). Our general goal is to extend the skillset of robots such that they become better, faster and safer workers. We will tackle this challenge by first investigating motion retargeting for robots. Motion retargeting is the problem of mapping motions from one robotic platform to another, with different configuration, size and number of degrees of freedom. It is a first step towards having humans train robots how to perform complex tasks through demonstrations.
The goal of this project is to develop a retargeting algorithm that tracks the motions of a human and maps it to our robotic platform, in simulation as well as to the physical robot.
The candidate will work with an existing codebase in C++ and a physical robot and extend its functionality to enable:
- motion tracking of human arms
- retargeting human motions to robotic platforms
- smooth and synchronized motions of the simulated and physical robot
The candidate will work with an existing codebase in C++ and a physical robot and extend its functionality to enable: - motion tracking of human arms - retargeting human motions to robotic platforms - smooth and synchronized motions of the simulated and physical robot
The goal of this project is to develop a retargeting algorithm that tracks the motions of a human and maps it to our robotic platform, in simulation as well as to the physical robot.
The goal of this project is to develop a retargeting algorithm that tracks the motions of a human and maps it to our robotic platform, in simulation as well as to the physical robot.