Abstact

This paper presents a real-time self-collision avoidance algorithm for a differentially driven mobile manip- ulator. When the manipulator and the mobile base are close to each other, attractive force is exerted on the mobile base for the manipulator to come close to the collision-free area in the workspace of the manipulator. The force is always generated even if the manipulator approaches the mobile base in any direction. In this respect, the proposed algorithm has a distinct advantage over existing methods which focus on simply pushing the links away from each other and do not consider the mobile base with non-holonomic constraints. The force and the resulting torque are converted to the desired accelerations of the mobile base and they construct the avoidance task. The task is smoothly inserted with a top priority to the controller based on hierarchical quadratic programming. The proposed algorithm was implemented on the differentially driven mobile base with the 7-DoFs robotic arm, and its performance was demonstrated during various experimental scenarios.

This paper has been selected as the Best Conference Paper Award in the conference!

PDF

PDF file

Multimedias

Award

See More

Please Visit HQP project and IROS 2019 for the self-collision avoidance algorithm.