Robust In-Hand Manipulation with Extrinsic Contacts

Thus, it is We can make very skillful use of various contacts desirable that a planning algorithm be robust to various (e.g., with the environment, our own body, etc.) to perform uncertainties like grasp center, extrinsic contact location, etc. complex manipulation. In a striking contrast, achieving such We present a method which can incorporate uncertainties in dexterous behavior for robots remains very challenging. Using several of the kinematic constraints to generate robust plans environmental contacts efficiently can provide additional for perform in-hand manipulation. This idea is also illustrated dexterity to robots while performing complex manipulation in Figure 1, where a naïve plan can easily lose contact with the [1]. However, the current generation of robotic systems environment due to uncertainty in the grasp location or the mostly avoid making contacts with their environment.