Desingularization of Flexible-Joint Parallel Robots

Abstract

Parallel robots possess a characteristic type of singularities, called type II or drive singularities, inside their workspace. In the neighborhood of these singularities, the inverse dynamics solution grows unboundedly and the robot becomes uncontrollable. There is growing literature on methods that enable parallel robots to pass through drive singularities. Most of this literature relies on dynamic models that presume rigid joints. However, the flexibility of the drive train elements should also be taken into account for high accuracy. In this paper, we propose a systematic trajectory planning method for enabling flexible-joint parallel robots to pass through drive singular configurations. Our method generates admissible polynomial trajectories of degree eleven. Four conditions are derived and incorporated into the method to prevent undesired back-and-forth motion of the endpoint. This ensures not only an efficient operation of the robot but also the avoidance of unintended multiple occurrences of the same singularity. The boundedness of the inverse dynamics solution is also guaranteed.