An efficient framework for multiple tasks in human-like robots

Human-like robots are required to simultaneously execute multiple tasks. A task-priority strategy plays an important role in implementing multiple tasks. Conventional task-priority strategies suffer from algorithmic singularity and large computational effort. Extended Operational Space (EXOS) formulation provides an algorithmic singularity free and computationally efficient framework for a single robot manipulator with two tasks. An extension to EXOS for multiple tasks, named Task-priority based EXOS (TPEXOS), has been proposed for whole-body control of human-like robots. TPEXOS augments constraint task spaces according to the order of priority without algorithmic singularities. The computational efficiency of TPEXOS excels other conventional task-priority strategies, whose efficacy and efficiency were demonstrated through simulation studies.