Device-free human sensing and localization in collaborative human-robot workspaces: a case study

Modern robot manufacturing is fostering the implementation of hybrid production systems characterized by human-robot cooperative tasks. Safety technologies for workers protection require advanced sensing capabilities and flexible solutions that are able to monitor the movements of the operator in proximity of moving robots. This paper proposes the use of wireless device-free localization (DFL) methods and architectures to detect and track a human worker in a cooperative human-robot industrial workspace. The DFL system is composed of groups of massively-interacting small, low-cost, embedded radio-frequency transceivers that perform received power measurements. These devices are anchored in fixed peripheral locations of the plant and provide localization of the worker, who peculiarly carries neither wireless active devices (device-free) nor specific tracking sensors (sensor-free sensing). Operator motion is, in fact, estimated by tracking the perturbations of the radio field induced by the human body, considering the effect of concurrently moving robot as non-stationary interference. The proposed localization and detection algorithm is based on the Jump Linear Markovian System - Interactive Multiple Model method and its positioning accuracy has been validated by experiments performed inside a robotic cell of an industrial test plant. The proposed DFL system has been implemented by employing IEEE 802.15.4 radio-frequency devices operating at 2.4 GHz and integrated into a software safety architecture. Finally, a software toolset has been designed to predict DFL accuracy, to verify experimental measurements and also to support the integration with pre-installed industrial sensors to increase the accuracy of the augmented system.