Assessing Latency Cascades in Human-Robot Collaboration Applications

Advancements in sensing technology and artificial intelligence have revolutionized industrial settings by introducing robots that work alongside humans, enhancing productivity and flexibility. However, ensuring safety in human-robot interactions has become more challenging. Established safety standards emphasize risk assessment, protective measures, and real-time monitoring systems, where safety complexities arise from intricate industrial interactions. The study focuses on "Speed and Separation Monitoring" (SSM), a collaborative type defined by ISO/TS 15066. This thesis aims to address unknowns within SSM, particularly on the parameter accounting for the robot system to respond to the operator's presence, crucial for decision-making on speed and separation limits. A proximity sensor was utilized to assess the overall delay of a classic industrial network between the sensing node for the operator detection (AI-based vision system) and the triggering of the safety node to the robot. The methodology was tested on a cohort of 23 subjects and evaluated under various lighting conditions. The work identified bottlenecks and the impact of each subsystem composing typical industrial control networks, highlighting the need for precise methodologies to assess latency as a critical factor in safety and productivity as sensing technology, collaborative robots and safety networks keep evolving.