The proliferation of robots in shared human environments demands robust safety in human–robot collaboration. In this article, the authors propose a Minimal-Distance Monitoring system aimed at augmenting safety within industrial robotic cells. It adopts a decentralized, multi-camera perception system to achieve precise, real-time pose estimation of human operators throughout the shared workspace. By fusing sensory inputs from heterogeneous viewpoints, the system enhances spatial perception, mitigates occlusion effects, and ensures robust detection of human postural configurations. Continuous computation of the minimal distance between the human model and robot’s links enables risk assessment during human-robot collaboration operations. This persistent monitoring mechanism not only enforces safety compliance but also facilitates the optimization of collaborative robotic workflows in contemporary industrial automation contexts.
Grajeda et al. (Thu,) studied this question.