Talk Title: Contact-Rich Manipulation: Bridging the gap between planning and control.
Abstract: Contact-rich manipulation scenarios where the robot seeks to make contact interactions with the environment in order to accomplish a given task, rather than avoiding it, are a great challenge in robotics. Among other factors, the contact interactions are difficult to model, introducing large uncertainty in the prediction of the motion; switching contacts introduces hybrid dynamics; and many contact interactions, such as non-prehensile manipulation, are underactuated. While control methods can make robots more reliable to these sort of forceful interactions, they usually lack the prediction capability to handle complex tasks. On the other hand, planning approaches are able to generate motions for contact-rich scenarios, but often fail to transfer to the physical robotic hardware. In this talk, I will argue that for achieving dexterous and agile manipulation capabilities in contact-rich scenarios, we have to carefully consider the interplay between both the control and planning aspects of the robot motion. To aid the discussion, at the end I will provide a counter example of the main thesis.