392248 ISY Project: Manipulation of Objects with Robotic Arms in Simulated Environments (Pj) (SoSe 2019)

In the project, you will train a human-like robot hand to manipulate objects in a simulated environment [1, 2] using general-purpose deep reinforcement learning algorithms [3], modern neural network architectures [4] and classic approaches [5, 6, 7, 8, 9]. You will experiment with adding touch sensors [10] to the hand and explore generalization of learned policies over different objects with and without tactile information.

[1] http://gym.openai.com/envs/#robotics
[2] https://drive.google.com/open?id=1J2H92AstGpcFYqjmVymMSRKI5xISdB1b
[3] https://github.com/openai/baselines
[4] https://github.com/atenpas/gpd (Grasp Pose Detection)
[5] https://youtu.be/-L4tCIGXKBE (Differentiable Physics)
[6] https://github.com/MarcToussaint/18-RSS-PhysicalManipulation
[7] https://moveit.ros.org
[8] https://youtu.be/0og1SaZYtRc (MoveIt)
[9] https://youtu.be/Gzt2UoxYfAQ (Contact-Invariant Optimization for Hand Manipulation)
[10] https://drive.google.com/open?id=1iGIbu00IPmI7IjmJrTUaYLIrKZzcgpdL (Touch)

Please note that the teams will be selected by the supervisors on the basis of short applications that students are expected to send to them. Registering to the project in the ekVV will only be regarded as expression of interest; it will not secure a team membership. Please get in touch with the supervisors for information on the application procedure.

Requirements for participation, required level

Required skills:

• Introduction to Neural Networks course or Advanced Neural Networks
• Python or C++ (>= 1 year)