Coupled recursive estimation for online interactive perception of articulated objects

We present online multi-modal perception systems for extracting kinematic and dynamic models of articulated objects from physical interactions with the environment. The systems rely on a RGB-D stream, contact wrenches, and proprioception. The proposed systems share an algorithmic foundation: they are based on an architecture of coupled recursive estimation processes. We present and advocate this architecture as a general, versatile, and robust solution for online interactive perception problems. We validate the architecture in extensive experiments to extract kinematic models interactively, varying the appearance, size, structure, and dynamic properties of objects for different tasks and under different environmental conditions. In addition, we experimentally show that the information acquired by the online perception systems enables robot manipulation of articulated objects. Furthermore, we discuss the relationship between the proposed architecture for robot perception and insights about biological perception systems.