ROS 2 Application to the Safety Monitoring of Collaborative Robots

This paper presents the application of a safety monitoring algorithm for collaborative robots (cobots). Cobots become increasingly common in various workplaces, such as manufacturing lines and even restaurant kitchens. The heavy-duty robots can assist human operators by handling heavy loads, reducing the physical strain of the job. However, safety concerns arise due to the close proximity and direct interaction between cobots and humans. To ensure real-time monitoring of potential accidents, this paper considers the use of a redundant fault detection algorithm. This algorithm relies on comparing outputs from redundant systems or sensor data with the actual robot's performance. Any significant discrepancy between these values signals a potential fault in the robot, triggering an accident alarm. In this study, the Gazebo simulator was employed as the redundant system. Gazebo is a widely-used, open-source 3D robotics simulator that integrates seamlessly with ROS (Robot Operating System). ROS 2, the next generation of ROS, serves as the middleware framework for building and executing robotic applications. The heavy-duty robot was modeled in Gazebo and controlled through ROS 2 nodes for communication and operation. A synthetic fault was introduced in one of the robot's joints, and another Gazebo simulation acted as the real system. Joint positions were monitored, and any deviation from normal operations signaled the presence of a fault. Consequently, integrating Gazebo with ROS 2 provides an effective analytical model to compare against real systems for safety monitoring purposes.