Optimal robot motion planning for human-robot collaboration considering power and force limiting

The operating mode power and force limiting protects the human operator during physical human-robot collaboration by limiting contact-related quantities. Using biomechanical injury criteria and contact models, thresholds for the energy transferred to the body part during potential contacts can be estimated for unconstrained contacts. In turn, robot velocity limits can be derived from the transferred energy thresholds and can subsequently be taken into account in motion planning and safety functions. A common assumption is to consider the effective robot mass as constant, which was shown to not always be conservative. In this paper, trajectories are optimized for a fixed path constraining the transferred energy calculated with the configuration- and direction-dependent robot mass instead. This yields optimal motions, satisfying transferred energy thresholds in cases where the approximation is non-conservative. For cases in which the constant mass approximation is conservative, cycle time reductions are achieved. The cycle time can be reduced further by optimizing trajectories without prescribing the path a-priori showing the potential towards increased motion efficiency for applications with power and force limiting.