SCIEN 2017: Professor Jeannette Bohg

Interactive Perception for Autonomous Robotic Manipulation

Abstract: Recent approaches in robotics follow the insight that perception is facilitated by interactivity with the environment. These approaches are subsumed under the term of Interactive Perception (IP). It provides the following benefits: (i) any type of purposeful interaction with the environment creates a rich, informative sensory signal that would otherwise not be present and (ii) any prior knowledge about the nature of the interaction supports the interpretation of the generated signal.
In this talk, I discuss the requirements for building a robotic system that is capable of interactive perception and can thereby leverage these aforementioned benefits. I instantiate such a system with an example in which real-time visual perception meets reactive motion generation. I demonstrate its capabilities in grasping and manipulation in challenging dynamic scenarios that go beyond commonly considered static scenarios.
As an example component of this system, I explain in depth the real-time visual tracking method that provides continuous feedback on the object and robot arm pose during manipulation. I also give an outlook on how this system can be used for autonomous, interactive learning of manipulation and grasping.

Biography: Jeannette Bohg is an Assistant Professor for Robotics at the Computer Science Department of Stanford University. Previously, she has been a group leaderat the Autonomous Motion Department, Max Planck Institute for Intelligent Systems in Tübingen, Germany. She holds a Diploma in Computer Science from the Technical University Dresden, Germany and a Master’s degree in Applied Information Technology from Chalmers in Gothenburg, Sweden. In 2012, she received her PhD from the Royal Institute of Technology (KTH) in Stockholm, Sweden.   Her research interest lies at the intersection between robotic manipulation, Computer Vision and Machine Learning. Specifically, she analyses how continuous, multi-modal sensory feedback can be incorporated by a robot to learn robust and dexterous manipulation capabilities in the presence of uncertainty and a dynamically changing environment.