Imitation learning for contact-rich manipulation

点击翻译按钮获取中文摘要

Currently, our goal is to extend the CTF IL framework [6] to handle contact-rich tasks byimproving the way object interaction occurs. The objective is to create a method that accountsfor force feedback as well as being closed-loop in order to adapt to errors arising during contactor from the visual feedback that can occur when approaching the object. Initially, we are focusingon scenarios where a new demonstration is provided for each new object, where providing arepresentative demonstration to solve the task is possible. Then, in the future we aim to exploremethods that build control policies based on 3D object models constructed directly from vision andtasks that require tactile feedback to be solved. Additionally, we aim to extend this to be able togeneralise demonstrations across similar objects and utilise prior knowledge obtained either in thereal-world or simulation to provide solutions for tasks were providing a clear enough demonstrationis not possible. Looking forward, the ambition is that by the end we can have a framework wherehumans can teach robots to solve complex, contact-rich manipulation tasks in the real-world, withonly a handful of demonstrations and improve from interactive human feedback.General research area: robotics