A COBOT-IMU Hand-Guiding System with Online Collision Avoidance in Null Space

In this study, we investigate the capability of human-robot interaction (HRI) using wearable sensor systems and a redundant collaborative robot (COBOT). An IMU suit records the motion of the human operator, whose left hand guides a redundant COBOT. The redundancy of the COBOT allows performing a desired hand guiding task with the redundant joints maximizing the distance from the operator to avoid contacts. In this work, we integrated the IMU suit into ros_control framework to achieve real-time capabilities. A motion reconstruction strategy based on reference pose calibration is used to estimate the body positions based on IMU attitude data. The minimum distance between operator and robot is then calculated and a null-space collision avoidance control strategy is implemented to avoid collision with the operator. A simple experimental test has been conducted to verify the proposed system and routine, which also confirms the IMUs’ capability of reconstructing the motion online.