A New Conversion Method to Evaluate the Hazard Potential of Collaborative Robots in Free Collisions

With the increasing importance of collaborative robots in industrial manufacturing, their economic efficiency is becoming more and more important. Today, one approach to prevent collaborative robots from injuring humans is to assure that the robot cannot exceed biomechanical limits in the event of an accidental collision or clamping. The ability of the robot to avoid collision forces beyond the limits must be validated with a biofidelic measurement device that mimics the biomechanical behavior of the human. For reliable use, the measurement devices must be attached to a rigid frame. Consequently, the test setup is solely able to simulate the contact dynamics and biomechanical consequences of a clamping contact in which the human body part cannot move. Free collisions that allow the human to move freely reduce the collision forces, but can only be evaluated with such a measurement device. This technical limitation leads to slower robots and thus a loss of productivity. The study presents a method that increases the efficiency of safe collaborative robots by adding a new validation procedure. The presented method incorporates a model-based conversion of measurements that enables safety experts to validate robots in free collisions. Data from experimental tests with a collaborative robot and a biofidelic measurement device show a good fit of the model-based prediction and thus confirm our approach. This new approach has great potential to increase the productivity of collaborative robots, since our method will allow them to move at faster but still safe speeds.