Architecting Hand-Guiding Collaborative Robots

Definitional Boundaries & Taxonomy

Hand-guiding collaborative robots (cobots) represent a specialized taxonomic branch within industrial robotics, defined by their intrinsic and extrinsic ability to safely share a workspace with human operators and physically interact with them in real-time. Unlike traditional industrial manipulators that operate strictly within isolated safety cells (fenced environments), hand-guiding cobots employ "kinesthetic teaching" or "lead-through programming". This paradigm allows the human operator to physically grasp the end-effector (or any link of the robotic arm) and move it through space to define waypoints, trajectories, and operational logic. Taxonomically, these robots fall under the ISO 10218-1/2 standard for Collaborative Operation, specifically operating in "Hand Guiding" (HG) and "Power and Force Limiting" (PFL) modes. The definitional boundary is marked by the presence of a "zero-gravity" or "gravity compensation" control state, where the robot actively nullifies its own mass and dynamic friction to feel entirely weightless to the human hand.

Historical Evolution

The conceptual origin of cobots dates back to 1996 with J. Edward Colgate and Michael Peshkin’s foundational patent at Northwestern University, describing an apparatus for direct physical interaction with a human. Early manifestations were entirely passive systems, utilizing continuously variable transmissions (CVTs) to steer mechanical power provided purely by the human worker. The leap to active hand-guiding occurred in the early 2000s with the KUKA LBR series (Leichtbauroboter) developed in collaboration with the German Aerospace Center (DLR). By integrating joint torque sensors, the LBR arm achieved unprecedented backdrivability. The paradigm exploded in 2008 when Universal Robots introduced the UR5, replacing expensive joint torque sensors with highly accurate motor current monitoring to achieve gravity compensation mathematically. This transition from purely physical compliance to software-driven virtual compliance catalyzed the modern era of democratic, easy-to-program robotics, shifting the programming interface from complex teach pendants to intuitive physical guidance.