Experimental evaluation of LoRa(WAN) in indoor and outdoor environments

Over the past years, Low Power Wide Area Networks (LPWANs) have become one of the fastest growing wireless technologies. These networks are ascribed a big role in the future of Internet of Things (IoT) applications and have gained a lot of interest, even though much practical details about them are still unknown. One such an LPWAN is LoRa, a proprietary spread spectrum modulation scheme. The spreading of the spectrum in LoRa modulation is achieved by generating signals in which the frequency linearly increases or decreases over time with a speed indicated as the spreading factor (SF). This thesis provides a theoretical and experimental evaluation of LoRa for both indoor and outdoor environments in order to determine LoRa’s suitability to provide a reliable solution for network connectivity. LoRa is proven an ideal technology for low energy consuming data transmissions over a long-range with low data rate requirements. Extensive experiments performed to evaluate LoRa performance have shown that it performance iss influenced by its environment. Indoor experiments have shown that LoRa’s performance is sensitive to both the location inside the building, as well as daily changes in the physical environment due to functional use of the office floor. Outdoor experiments have shown that mobility and the speed of the movement do not negatively influence performance. The distance and type of environment do play a role with less dense environments in terms of physical objects are performing the best. Empirical results have shown that LoRa’s outdoor performance can be increased by lowering the transmission data rate. Simultaneous LoRa transmissions on the same frequency and SF can cause collision-induced packet loss depending on the timing of the transmissions. During such an event, there are conditions under which the stronger of the two signals can be successfully received due to the Capture Effect. Most of the time, however, at least one of the signals is lost. The ability to provide reliable network connectivity in a dense IoT environment is examined by means of a simulator, whose behavior is based on the empirical results found in this thesis. Subsequent simulations of environments in which a large number of LoRa end-devices are present show that a few thousand end-devices can already induce performance degradation. Several solutions to cope with collision-induced packet loss or to reduce the chance that it will happen are examined. A combination of (i) decreased base station coverage, (ii) smaller data packets, and (iii) more transmission channels is found to be the best solution to prevent network congestion. Ultimately, the reliability of a LoRa network depends on the density of LoRa devices in the neighborhood and the physical characteristics of the environment in which end-devices are placed.