Towards Shared Autonomy Applications using Whole-body Control Formulations of Locomanipulation

Abstract

While widely studied in robotics for decades, mobile manipulation has recently seen a surge in interest for industrial applications due to increasing demands on flexibility and agility alongside productivity, particularly in small and medium enterprises. However, most mobile manipulation solutions frequently decouple the navigation from the manipulation problem effectively performing fixed-base manipulation using a repositionable manipulator. This is not only inefficient, but moreover limits the range of applications and disregards the inherent redundancy of a mobile manipulation system. In this work, we introduce a high-performance omnidirectional mobile manipulation platform with integrated whole-body control, real-time collision-free whole-body motion planning, and perception. We demonstrate its capability along with application scenarios on technical demonstrators involving moving manipulation targets as well as whole-body manipulation in simulation and hardware experiments. Finally, we deploy and evaluate our solution in field trials on an industrial oil and gas training facility on a sensor placement and manipulation task.