Human Intention Prediction in Human-Robot Collaborative Tasks

Enabling the robot to predict human intentions in human-robot collaborative hand-over tasks is a challenging but important issue to address. We develop a novel and effective teaching-learning-prediction (TLP) model for the robot to online learn from natural multi-modal human demonstrations during human-robot hand-overs and then predict human hand-over intentions using human wearable sensing information. The human could program the robot using partial demonstrations according to the updated tasks and his/her personal hand-over preferences, and the robot can online leverage its learned strategy to actively predict human hand-over intentions and assist the human in collaborative tasks. We evaluate the approach through experiments.