COLMAG - A Novel Human-Aware Motion Framework Exploiting Magnetic Sensing in Collaborative Robotics

Abstract The integration of human-friendly robots into dynamic, shared environments has led to advancements in sensing and control strategies to ensure safe Human-Robot Interaction. Existing solutions rely on vision-based systems or capacitive skins for human motion tracking and collision avoidance, but these approaches can face limitations including occlusions, sensitivity to environmental conditions, or restricted detection range. Here, we propose COLMAG, a novel approach that merges COLlaborative robotics with MAGnetic sensing to enhance collaboration in physical Human-Robot Interaction. The system incorporates a magnetic marker worn by the human operator, whose position is tracked by a sensor array mounted on the endeffector of a robotic manipulator. A computationally efficient algorithm decodes the magnetic field data, enabling precise localization and intent recognition. We present a multi-modal control architecture and experimentally validated it in three representative motion generation modalities: collision avoidance, safety, and contactless guidance, using velocity fields in task space. Results highlight the potential of magnetic sensing to guarantee robot responsiveness and adaptability, offering a robust complementary solution to traditional sensing methods and opening new avenues for research in safe, intuitive, and contactless human-robot collaboration.