Skill learning for assistive robotics

Health-care and assistive robotics is the emerging industrial in the coming decades due to the increase of the elderly population. Assistive robots are demanded to assist aged people in performing various tasks. Thus, it is essential to endow assistive robots with the ability to adapt learned skills for performing different tasks. This paper proposes a skill learning framework to enable assistive robots to utilize basic skills for accomplishing various tasks. In the proposed framework, the major difference from the previous work is its capability of using basic skills to accomplish various tasks and each basic skill is composed of stimulus-response units (SRU). The proposed skill-learning framework can be realized by acquiring basic skills and utilizing them for assisting people to perform some of daily tasks. Computer simulations and experiments with a Pioneer 3-DX robot are conducted to validate the learning capability of the proposed skill-learning framework.