Smart Robot Monitoring Using Edge and Cloud Computing

What is it about?

This work is about improving how we check if a robot’s software is working correctly while it’s running. Robots often do complex tasks and need software that behaves as expected. One way to check this is with "assertion-based verification," which watches what the robot is doing and makes sure it follows the rules. However, this method can use up a lot of computing power, especially on robots with limited hardware. As a result, it can slow the system down and even cause the robot to miss important timing deadlines, which may lead to failures during operation. To solve this, we built a system that splits the monitoring work between the robot’s local computer (the "edge") and more powerful cloud computers. The system automatically moves parts of the monitoring task back and forth depending on the robot’s current workload and available resources. Everything is packaged into easy-to-deploy software units using Docker and works with ROS, a common framework for robotic systems. We tested our solution with simulated benchmarks and a real-world factory robot to show how well it performs under different conditions.

Featured Image

Why is it important?

As robots are used more in factories, hospitals, and homes, making sure their software runs safely and correctly becomes crucial. Our method helps catch problems without slowing down the robot, by using extra computing power from the cloud when needed. This makes it easier to build reliable robots that can be safely used in real-world environments, even when their onboard computers are limited.

Perspectives