Trunk Robot

Many real-life tasks ask for robots that are flexible and dexterous – from cautiously helping elderly, to safely repairing spacecraft. We anticipate that the next generation of robots will embody a continuum design, enabling them to assume a virtually limitless range of shapes and movements. Albeit promising from an application point of view, this flexibility also brings with it unique challenges in accurately modeling such systems. Their complexity, nonlinearity, and high dimensionality make it difficult to obtain a compact representation of their physics suitable for control. The Autonomous Systems Lab (ASL) has developed algorithms to judiciously find such models, enabling precise control and guarantees on safety and robustness.

This demo aims to showcase such techniques on our in-house developed robot, namely the ASL Trunk Robot. This is a low cost, open-source desktop continuum robot platform, equipped with a gripper to perform manipulation tasks. To demonstrate the usefulness of such soft robots, the user can interact with the robot by controlling it using an Apple Vision Pro, for which we developed a custom app. The user hands the robot tools and waste items, which can then be placed in the corresponding containers via hand gestures. The arm, being flexible, is unable to endanger the user, allowing closer human-robot interaction. In the future, such tasks will be done fully autonomously, by leveraging data collected from these demonstrations.