Aalto university presented the current development on their work related to the MirrorLabs project in the 26th International Conference on Emerging Technologies and Factory Automation (ETFA 2021) (https://2021.ieee-etfa.org/ )
Nowadays, Automated Guided Vehicles (AGVs), also known as Autonomous Mobile Robots (AMR) are an important part of the manufacturing sector. Visualizing an object that moves autonomously on the shop floors might be a complicated task for current SCADA and HMI systems. For that reason, the Aalto Factory of the Future team implemented a visualization demonstrator leveraged on Gazebo, a popular open simulation tool produced and maintained by the Open Robotics Foundation, which is also based on the ROS environment. The approach is based on using information produced by the edge robot controller (onboard computer), such as the path plan and localization, which is published over a ROS topic and is accessible to Gazebo and its web component GzWeb for web visualization purposes. The gazebo components are hosted in a Linux virtual machine deployed as a service on the Azure cloud (West Europe - Amsterdam). Then, given a domain provided by the cloud service, it is possible to access the visualization environment from any compatible platform that supports the web browsing technologies used by GzWeb, for instance, a phone, a tablet, or a computer regardless of its operating system.
Until now, everything sounds great, but several technical challenges might affect the user experience, one of the most important is the quality of service for the communication since multiple wired and wireless networks link the data from the robot to the cloud and client. This problem is addressed by the presentation of the Aalto team in ETFA 2021, which showcases the progress of the efforts to develop a methodology to enhance the real-time visualization of remote-controlled AGVs.
The robot trajectory visualization is shifted from a real-time velocity report to the progressive path plan, completely based on a pose list in the coordinate system used for localization. Then, the system will visualize the 3D representation smoothly through a web browser, while the robot navigates and reports its localization, which comes with a certain communication delay to the visualization platform. The delay data is constantly monitored and based on a delay prediction system, it is possible to provide early alerts about communication malfunctions or uncertainty on the dynamic visualization accuracy.
The expected outcome of this work is improving the user experience of remote real-time visualization consumers who are monitoring the behavior of cyber-physical systems in agile manufacturing. In the future, the team will focus on improving the methodology behavior, integrate it with PLC-based systems, and migrate the visualization platform to Ignition, a new Open Robotics Foundation platform taking over Gazebo.
