Opportunistic sensing in beyond-5G networks: the opportunities of transformative computing

Transformative computing, or cross-fertilization of computing and communication technologies, has stimulated a lot of research interests by leveraging existing network and hardware to build new sensing infrastructures on top of traditional communication services. Beyond-5G networks are thus expected to blur the boundaries between wireless systems, core networking, Artificial Intelligence (AI) and computing architectures. They are aimed to adapt to, and interact with, the surrounding environment, leveraging a wide range of systems, devices, tools and techniques. A fundamental, and often unspoken, assumption to such an evolution is that networks will be able to accurately sense the environment itself. Among sensing approaches, opportunistic sensing has emerged to prominence, thanks to its ability to exploit the stray radio-frequency radiation associated with the wireless devices. In this chapter, we describe how distributed radio sensing techniques and machine learning tools can be coupled with 5G architectures based on the multi-access edge computing (MEC) paradigm, in order to improve the efficiency and effectiveness of opportunistic sensing. We quantify the benefit of adopting a MEC-based architecture within a distributed processing and machine learning framework with reference to two prominent opportunistic sensing applications, namely, home automation and collaborative robotics.