Achievable push-manipulation for complex passive mobile objects using past experience

Majority of the methods proposed for the problem of push planning and manipulation deal with objects that have quasistatic properties and primitive geometric shapes, yet they usually use complex physics modeling for the manipulated object as well as the manipulator. Our proposed approach is an experience-based one, where the mobile robot experiments with pushable complex real world objects to observe and memorize their motion characteristics together with the associated uncertainties in response to various pushing actions. This incrementally built experience is then used for constructing push plans based solely on the objects’ predicted future trajectories without any need for object-specific physics or contact modeling. We modify the RRT algorithm in such a way to use the recalled robot and object trajectories as building blocks to generate achievable and collisionfree push plans that reliably transport the object to a desired 3 DoF pose. We test our method in a realistic 3D simulation environment as well as in a physical setup, where a variety of pushable objects with freely rolling caster wheels need to be navigated among obstacles to reach their desired final poses. Our experiments demonstrate safe transportation and successful placement of the objects.