2 year full-time postdoc position in impact-aware robot manipulation

Are you fascinated by robotic manipulation and would like to contribute to a step-change in robotic dexterity? At the TU/e, we are looking for a passionate and skilled researcher with proven experience in robotic manipulation and perception, willing to join a lively European project consortium focusing on impact-aware manipulation (https://www.i-am-project.eu ).

JobDescription

The I.AM. project is enriching traditional pick-and-place robot technology by developing human-like grab-and-toss capabilities. Driven by real use-case scenarios from the logistic industry, we are developing modeling, learning, planning, perception, and control methods to exploits intentional collisions for manipulation.

The Dynamics and Control Robotics and Perception group of the TU/e is coordinator of the I.AM. project and our role in this project concerns the collection and long-term storage of robot-object-environment impact data, the development of predictive models for post-impact state forecasting, and the foundation and validation of impact-aware robot control. Besides these key activities, the group actively support the physical realization of the three validation scenarios of the project (TOSS, BOX, and GRAB), focusing of having robot tossing objects on conveyor belts, densily stacking objects in tote containers, and swiftly grabbing boxes from a pallet, respectively.

Your task will include:

• perform top level research, publish, and present your work at national and international gatherings/conferences
• help coordinating research activities within the I.AM. project, by supporting the coaching of PhD and MSc students
• support the ongoing hardware/software integration efforts, started 2 years ago, combining Algoryx Dynamics (https://www.algoryx.se/agx-dynamics/ ) and QP robot control (https://jrl-umi3218.github.io/mc_rtc/ ), used within the I.AM. consortium to develop impact-aware modeling, learning, estimation, and control
• assist in teaching activities (4 hours/week practicals, plus preparation, on standard robot modelign and control; one academic quarter/year)

The research area is that of impact-aware robotics (cf. https://impact-aware-robotics.gitlab.io/IROS2021workshop/ ) and therefore it is expected that you have strong interest and relevant experience in specific topics such as

• control of robots in physical interaction (e.g., impedance control, joint torque control)
• modeling, control, external-wrench estimation of real robots with flexible joints, in contact (e.g., impedance control, QP robot control, mometum observer) and/or impact scenarios (cf. https://hal.archives-ouvertes.fr/hal-02434909 )
• data collection and storage of robot-object-environment physical interaction (the following database is being created at the TU/e https://impact-aware-robotics-database.tue.nl/ )

Two labs are available for the project purpose: (a) the robotics lab of the TU/e, provided in particular with a 12 cameras Optitrack mocap system, 2 panda arms, and VR HMD and controllers for teleoperation; (b) a joint industry-university Innovation Lab on TU/e campus, provided with an UR10 and 1 Panda robot arms, industrial vacuum grippers, 360Hz 6-camera motion capture system, and an industrial conveyor belt.