 |
 |
The integration of cloud computing technologies with robotics has opened new frontiers in the field of distributed robotic systems. As the Internet of Things (IoT) and embedded systems continue to proliferate, there is an increasing need to satisfy computational requirements at the proximity of these devices and cyber-physical systems. This demand has pushed the boundaries of computation beyond centralized cloud infrastructure to include edge computing layers. This evolution has given rise to what is known as the device-cloud continuum, a paradigm that incorporates end devices, robots, and cyber-physical systems as part of a seamless computational environment.
The device-cloud continuum leverages computational resources across a spectrum, from end devices through edge computing layers, all the way to cloud infrastructure. This approach enhances the capabilities of robotic applications by allowing flexible allocation of computational tasks. For instance, time-sensitive operations can be performed closer to the robot, while more resource-intensive computations can be offloaded to powerful cloud servers. By integrating robots as active participants in this continuum, rather than mere end-points, we can achieve a more efficient and responsive distributed robotic system.
This thesis discusses a clustering solution aimed at optimizing throughput while maintaining system performance in a Robot Operating System 2.0 (ROS2) distributed environment. The methods studied take another step forward towards robust and scalable scheduling of distributed robotics architectures, with the goal of enabling efficient and dynamic service placement in the device-cloud continuum.