Design and control of a novel omnidirectional dynamically balancing platform for remote inspection of confined and cluttered environments

Remote inspection is a long standing field of interest for robotics researchers, in which robots are used to undertake inspection tasks in environments too hazardous or inaccessible to be directly entered by a human. Recent advances in gridscale battery storage have created a new set of unique hazardous indoor spaces with characteristics unsuitable for the deployment of existing teleoperated inspection robots. This paper outlines the problems encountered in these new environments, analyses existing approaches to robotic platform design, and proposes a better suited novel platform design, based on a dynamically balancing arrangement of Mecanum wheels. Its inverse kinematic and dynamics models are developed, a proof of concept prototype is constructed, and a constrained predictive controller is derived from the developed model. Experimental results demonstrate the efficacy of this new concept.