Human-aware task and motion planning for efficient human-robot collaboration

In human-robot collaboration (HRC), the variability of the human behavior complicates the deployment of robust task and motion plans, and continuous update of plans are often required to correct the task execution. We present a control-based approach to achieve robustness in the execution despite the high time variability of human and robot tasks. The proposed approach consists of two layers: task planning considers high-level operations without taking into account their motion properties; action planning optimizes the execution of high-level operations based on current human state and geometric reasoning. The result is a hierarchical structure where the lower level provides feedback on the feasibility and the upper level uses this feedback to (re)-optimize the process plan only when needed. The method is demonstrated in an industrial case study where a robot and a human worker work together to assemble a mosaic.