Physics-aware collaborative assembly with mobile robots

The ability to collaboratively build functional artifacts that are otherwise impossible to achieve alone is a crucial part of human intelligence. This project explores ways to endow robots with such intelligence in the context of assembly, so that they could use physical laws to reason and coordinate collective efforts to achieve a common construction goal. Our goal is to enable collaborative behaviours to emerge compositionally from low-level primitive skill and an understanding of the physical behavior of the assembly, by combining state-of-the-art Task And Motion Planning (TAMP) and physics simulation techniques.

The ability to collaboratively build functional artifacts that are otherwise impossible to achieve alone is a crucial part of human intelligence. This project explores ways to endow robots with such intelligence in the context of assembly, so that they could use physical laws to reason and coordinate collective efforts to achieve a common construction goal. Our goal is to enable collaborative behaviours to emerge compositionally from low-level primitive skill and an understanding of the physical behavior of the assembly, by combining state-of-the-art Task And Motion Planning (TAMP) and physics simulation techniques.

Building on the mentor’s research on Task And Motion Planning, the student will work on deploying a long-horizon assembly plan to one or several mobile robots in the real world so that a large-scale structure can be built with precision. We will use a motion capture system for localization and object tracking, and focus on the integration of the tracking system, planning, and control so that the robot could adjust to the tracking inputs to ensure accuracy. The student will work on the implementation and integration of the following components: - Trajectory planning for the differential-drive mobile base for smooth, collision-free motion. - Control strategy for the mobile base to follow the planned trajectory - Pose estimation of the built structure and the mobile base using the motion capture system - Augumented reality visualization as a qualitative verification of the control and tracking performance. - Integration of the above components to have one or several robots build a tetrahedral structure with wooden bars and swivel couplers.

Building on the mentor’s research on Task And Motion Planning, the student will work on deploying a long-horizon assembly plan to one or several mobile robots in the real world so that a large-scale structure can be built with precision. We will use a motion capture system for localization and object tracking, and focus on the integration of the tracking system, planning, and control so that the robot could adjust to the tracking inputs to ensure accuracy.

The student will work on the implementation and integration of the following components:

- Trajectory planning for the differential-drive mobile base for smooth, collision-free motion.

- Control strategy for the mobile base to follow the planned trajectory

- Pose estimation of the built structure and the mobile base using the motion capture system

- Augumented reality visualization as a qualitative verification of the control and tracking performance.

- Integration of the above components to have one or several robots build a tetrahedral structure with wooden bars and swivel couplers.

This is a paid summer research assistant position open only to ETH Zurich students. The student will be paid according to the ETHZ guidelines for research assistants. The student is expected to start as soon as possible and work full-time (40 hours per week) for 1 to 1.5 months until September.

This is a paid summer research assistant position open only to ETH Zurich students. The student will be paid according to the ETHZ guidelines for research assistants. The student is expected to start as soon as possible and work full-time (40 hours per week) for 1 to 1.5 months until September.