Abstract
We demonstrate manipulation of objects using the dynamics of a rope-like structure attached to a mobile robot as a passive tail. Three challenges arise in modeling and planning: the physics involved is nontrivial, the tail is underactuated, and motions of the object are nondeterministic. For such systems, some actions are well characterized by a simplified motion model (e.g., for dragging objects), but we resort to data-driven methods for others (e.g., striking motions). A sampling-based motion planner, adapted to deal with nondeterministic object motions, is used to optimize motion sequences based on a specified preference over a set of objectives, such as execution time, navigation cost, or collision likelihood. Experiments show that a robot with a passive tail can manipulate cylindrical objects with (quasi-static) dragging, dynamic striking motions, and combinations thereof. The method produces solutions that suit diverse preferences effectively, and we analyze the complementary nature of dynamic and quasi-static motions, showing that there exist regimes where transitions between the two are indeed desirable, as reflected in the plans produced.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have