A major theme of this fellowship project was to investigate wireless technologies for low power transmission over a long-range thus facilitating the operation of efficient wireless sensor networks. In collaboration with the host institution, the University of Cyprus (UCY), the Marie Curie fellow has led the investigation obtaining several significant results in this area that have been openly disseminated and published in flagship engineering conferences and journals.
On the brink of the IoT revolution, different low-power wide area network technologies are already competing for market share and scalability. LoRa networks is one such technology that has been deployed all over the world and is a major enabler for massive connectivity applications such as smart metering, agriculture, supply chain, and logistics due to the long-range, low cost, and low power features of LoRa. Thus, much of the work carried out in the project has used the LoRa network architecture as a starting point and use case.
The first main result of the project leveraged stochastic geometry tools to better understand how one should model the uplink wireless coverage in a multi-cell LoRa network while taking under consideration both the effects of co-spreading factor interference and the spatial diversity afforded by the LoRa network server. To that end, a tractable model that captures these two system peculiarities was proposed and later used to mathematically show that gateway densification has an overall positive effect on network coverage thus supporting the scalability of LoRa networks.
Building on this first result, more realistic non-uniform LoRa network deployments were also considered and mathematically investigated than what is usually the case in the state-of-the-art literature. Specifically, the question of how non-uniform deployments affect network coverage was studied analytically for the first time. To that end, the fellow and his UCY collaborators suggested optimal deployment strategies and uplink random access transmission schemes to improve network performance. It was found that concave deployments of LoRa end-devices with a sub-linear spread of random-access inter-transmission times provided a more optimal network coverage performance.
Energy harvesting technologies that extract energy from ambient sources have enabled the battery-less operation of many small wireless sensors. Therefore, the final main result of this project merged LoRa wireless communication systems with ambient energy harvesting sources and mathematically modeled device and network performance using tools from stochastic geometry and Markov analysis. To that end, the steady-state distributions of the capacitor voltage were derived, and the outage probability due to co-spreading factor interference at the LoRa gateway was calculated. This level of engineering insight enabled the fellow to propose efficient operating modes such as adaptive charging time schemes that can mitigate energy outage events and extend indefinitely the operation of the network while also maximizing the quality of service.
In addition to low power wireless communications, the project also investigated wireless localization challenges in collaboration with colleagues at the University of California Davis (USA), the University of Bristol (UK), the University of Nottingham (UK), the Paris-Saclay University (France), and Toshiba Corporation (Japan). Finally, in collaboration with researchers in the UK, the project also investigated and demonstrated wireless power transfer using ultrasound to remotely power all sorts of low-power devices, particularly in the context of human-computer interaction.
