Singularity Avoidance for Cart-Mounted Hand-Guided Collaborative Robots: A Variational Approach

Most collaborative robots (cobots) can be taught by hand guiding: essentially, manually jogging the robot, an operator teaches some configurations to employed as via points. Based on those points, Cartesian end-effector trajectories such straight lines, circular arcs or splines are then constructed. Such methods can, in principle, for cart-mounted cobots (i.e., when involves one two linear axes, besides cobot axes). However, applications, sole imposition of points space is not sufficient. On contrary, albeit overall system redundant, (i) must reached at joint configurations, and (ii) undesirable singularity (and near-singularity) conditions avoided. The naive approach, consisting setting cart trajectory beforehand (for instance, imposing a linear-in-time motion law that crosses configurations), satisfies first need, but does guarantee satisfaction second. Here, we propose approach novel strategy decoupling planning robot joints, variational technique computing former singularity-aware fashion, ensuring avoidance class workspace near-singularity configurations.