6G-RESCUE

Edge computing is set to accelerate innovation leading to new opportunities as it becomes an essential part of digital transformation across industries as more devices and applications demand low latency, real-time processing, and enhanced security. The 6G capabilities, such as network slicing, virtualisation and communication are of the great importance, but these solutions also require support for ML model training and deployment to the end devices.

The heterogeneity of the edge devices, their limited resources such as processing power and energy, calls for the dedicated services for fast ML model updates, model retraining and optimisation for embedded devices. Therefore, the edge component for efficient model serving and data processing for fast system reaction should be supported by the infrastructure. In the project, we will provide an extension enabling efficient MLOps operations on the edge as an addition to the core cloud services.

One of the notable examples would be hospitality and cleaning robots, which are automated machines and robotic systems designed to assist or enhance services in the hospitality industry, which includes hotels, restaurants, resorts, and other guest-centric businesses. In addition to their regular functions, they can be instrumental in disaster management. During events like natural disasters or security incidents, these robots can operate as support tools, helping locate people, detect suspicious objects, and provide essential information from their sensors and cameras.

For such a scenario, the functionality of the robots and the communication requirements have to be promptly reconfigured and updated. The object detection models should be fine-tuned with the images of lost people and deployed on the robots, while the video streams from the cameras should be prioritised to deliver life-saving information to crisis management centres.