Spectrum optimisation for low power wide area network utilisation in smart cities

Abstract

With the advent of the fourth industrial revolution, demand for spectrum by wireless networks has continued to increase exponentially. As the unlicensed band is a free-to-use spectrum, many networks including LPWANs gain access resulting in overcrowding and congestion. This research focuses on the optimisation of unlicensed spectrum occupied by LPWANs at 868MHz and 915MHz using the Optimised Frequency Hopping Model. The study focuses on these frequencies because these are where most LPWANs are found. The major challenge faced by these unlicensed spectra will continue to impact negatively on the network performance. For instance, there is a noise floor called the ambient noise which is a background noise that is present in the unlicensed spectrum due to signals of different wireless devices making use of these frequencies. On the other hand, Sigfox and LoRa; two of the most popular LPWANs operate at Sub-GHz frequencies of 868MHz and 915MHz in Europe and America. These unlicensed spectrum frequencies are susceptible to different problems like interference due to coexistence of networks and ambient noise. As long as it is a free-to-use spectrum, more devices will transmit at these frequencies below 1GHz causing disruptions to the services of LPWANs. It is therefore imperative to optimise the unlicensed spectrum in order to provide the ‘city of the future’ with seamless connectivity. Models like Open Sharing Model (OSM), Hierarchical Access Model (HAM), Dynamic Spectrum Sharing Model (DSSM) and Cooperative Spectrum Sharing Model (CSSM) were considered and tested. Their performances in the unlicensed spectrum at the frequencies of 868MHz and 915MHz showed little improvement against ambient noise and interference when compared to the proposed Optimised Frequency Hopping Model. Furthermore, the proposed Optimised Frequency Hopping Model showed a better performance against the troublesome ambient noise and interference, the two major problems associated with an overcrowded unlicensed spectrum. Simulations were carried out using MATLAB and results showed that the proposed method reduces interference at frequencies of 868MHz and 915MHz to the barest minimum. Hence, there is a decrease in latency, increase in channel capacity, high throughput, reduction in degradation of performance and higher quality of service QoS.