Using tactile sensors to estimate care receiver position on dual arms of robot

Abstract

Caregivers in nursing-care environments develop strain injuries while transferring patients from a bed to a wheelchair. We have developed a nursing-care robot, RIBA, which can perform transfer tasks using its two arms. RIBA must lift the appropriate part of the care receiver's body and transfer the care receiver between the bed and the wheelchair without inflicting pain or discomfort. To accomplish this, RIBA must know the care receiver's position in its arms when the arms are between the bed and the care receiver. To detect the points of contact between the care receiver and the robotic arms, the position of the care receiver on the robotic arms was estimated by using tactile sensors on the robotic arms as well as a machine learning algorithm. The results show that the caregiver position was estimated with 97% accuracy by our method.