Project title: 6D-Pose Estimation for Dexterous Manipulation

Description:
The research focuses on the shared-autonomy framework for robotic manipulation which combines 6D object pose estimation and trajectory planning subsystems for autonomous operation accompanied by assistance of human operator for complex and non-determined manipulation tasks. The proposed system is able to predict the 6D poses of objects, the level of internal liquid, the geometric location of the upper neck of the vessels and autonomously operate with the objects. Unlike full telemanipulation mode, it provides more robust and efficient trajectory as well as gripping functions. In addition, for more intuitive use, the digital twin of the robot is combined with feedback from visual and tactile channels in a virtual environment. As the next stage of system development, it is planned to apply Imitation learning algorithms using 6D object poses as key characteristics of the state of the environment, as well as information obtained from tactile sensors. The telemanipulation mode will be used to collect trajectories from demonstrations for training the model. The proposed framework will be useful for automating non-trivial manipulation tasks that require high accuracy and repeatability.

Candidates requirements:
➔ basic programming skills in Python or ROS/Gazebo

Supervisor: Associate Professor Dzmitry Tsetserukou

Internship duration: 1-2 months

The start date of the internship: July 12

End day: August 31

Contact person: Dzmitry Tsetserukou d.tsetserukou@skoltech.ru