Review of vision-based safety systems for human-robot collaboration

Safety in human-robot collaborative manufacturing is ensured by various types of safety systems that help to avoid collisions and limit impact forces to the acceptable level in a case of collisions. Recently, active vision-based safety systems have gained momentum due to their affordable price (e.g. off-the-shelf RGB or depth cameras and projectors), flexible installation and easy tailoring. However, these systems have not been widely adopted and standardized. The vision-based commercial products can only be found in a limited number. In this work, we review the recent methods in vision-based technologies applied in human-robot interaction and/or collaboration scenarios, and provide a technology analysis of these. The aim of this review is to provide a comparative analysis of the current readiness level of vision based safety systems for industrial requirements and highlight the important components that are missing. The factors that are analysed as such are use case flexibility, system speed and level of collaboration.