| Abstract: |
Human-Robot Interaction has emerged as a critical area of focus in modern robotics, particularly in industrial applications, where collaborative robots are increasingly employed. Various types of human-robot interaction exist — cell, co-existence, synchronization, cooperation, and collaboration — each involving different degrees of dependency between humans and collaborative robots, along with associated safety hazards. In our current work, we develop a framework relating safety hazards and mitigation strategies with interaction and dependency types. This will allow structured safety assessment of human-robot interactions.
Initially, precautions due to safety were implemented by confining robots within cages. This approach was necessary due to the mechanical advantages robots possess over humans, which, if not properly managed, could pose significant risks to both humans and the surrounding workspace. However, the introduction of human-robot interaction (HRI) in manufacturing has created a growing need to relax these restrictive measures. This transition requires advanced safety measures and comprehensive analyses to establish more precise and tailored safety mechanisms.
As the need for early safety planning becomes critical in the context of HRI, model-based engineering approaches, which have demonstrated their utility and benefits in industrial automation systems, can serve as valuable tools. Conceptual models, often depicted graphically, offer a formal representation of the modeling domain, capturing both static and dynamic aspects that are essential for the analysis, design and implementation of robotic systems.
In previous work, we proposed a model-based approach to relate requirements to corresponding safety hazards and countermeasures. This approach emphasizes the identification of hazards not only within individual systems but also arising from collaboration scenarios, contributing to extensive safety analysis during the early development stages of collaborative robotic systems.
In our current research, we build upon this approach. As safety considerations vary depending on the degree of interdependence between humans and robots, we identified dependency types that exist for HRI. By relating interaction types and dependency types, we are able to relate safety hazards to either the interaction, the dependency, or a combination. |