Design of a UAV-Based Videosurveillance System with Tactile Internet Constraints in a 5G Ecosystem

Abstract

One of the main goals of the new fifth generation of cellular networks (5G) is to provide the programmability, flexibility, and modularity that are required to create multiple logical networks, called network slices, each tailored for a given vertical use case, on top of a common network infrastructure. A particular network slice that is expected to be able to create a plethora of new opportunities and applications, is the Tactile Internet, aiming at guaranteeing e2e delays not greater than 1 ms to its users. Matching this very severe requirement does not only depend on the length of the physical path that information should follow flowing from sensors to actuators, but also on the size of each information. This is the case of a video-surveillance system that uses Unmanned Aerial Vehicles (UAV) to capture high-resolution images with a frame rate of thousands frames per second. Considering this kind of scenarios, this paper proposes a Tactile Internet architecture for video-surveillance systems. An analytical model based on Markov chains is defined to evaluate delays and job loss probability in order to allow designers to dimension system elements to be compliant with the Tactile Internet requirements.