Generating Motion for Adaptive Robots

The aspiration of modern robotics is to achieve a level of adaptability, robustness and safety that will allow the wide deployment of robots in unstructured domains and everyday human spaces. This requires progress at multiple components of robotics, from mechanisms, to sensing as well as decision-making and reasoning. This talk starts from robot planning algorithms, which achieve asymptotic optimality for systems with significant dynamics. It shows that these methods can be integrated with machine learning models and used for controlling new, soft robotic mechanisms that have been developed either for safe, effective grasping and manipulation, such as soft adaptive hands, or locomotion over uneven terrains, such as tensegrity robots. The talk will then proceed to describe perception methods, which can identify and robustly track objects without the need for cumbersome human annotation, and which allow the effective application of planning and control methods for adaptive manipulation, such as robotic packing and tight insertion for assembly. The talk will conclude with a brief discussion on how this progress empowers the next step in the field involving the development of robots that are socially cognizant and can be safely and effectively integrated into our society.