Dynamic singularity avoidance for parallel manipulators using a task-priority based control scheme

This paper proposes a control scheme that allows parallel manipulators to overcome a major inherent challenge, namely, avoiding singularities occurring inside their workspaces. The scheme lets a parallel manipulator follow a path that originally contains such singularities. By means of a suitably chosen artificial potential function arising out of the proximity to the singularities, the behaviour of the manipulator is modified dynamically, enabling it to push away the singularities in its path, without deviating from the path itself. Such behaviour is obtained, however, at the cost of one of the DoF of the manipulator. Nevertheless, various application scenarios involving axi-symmetric tools, e.g., industrial operations such as welding and gas cutting, can afford such a trade-off, and thereby benefit by eliminating the requirement of checking each path for singularities a priori. An additional benefit is the increase in the effective size of the workspaces due to the avoidance of internal singularities. The theory developed is applicable to any parallel manipulator, in principle. In this paper, it has been applied to a three- DoF planar 3-RRR manipulator. The experimental results clearly demonstrate the efficacy of the proposed scheme.