An evaluation of LoRa transmission parameter selection for indoor positioning

Sammanfattning: The purpose of this thesis is to investigate how the LoRa modulation configuration affects precision and accuracy when using LoRa for indoor positioning in non-line-of-sight conditions. The aim is to research if an optimum combination of spreading factor, bandwidth, and code rate factor can be found to result in the best possible positioning accuracy and precision, under certain predefined conditions. An experiment was conducted, where quantitative data was collected from an experimental setup. The experiment consisted of two testbeds which also included an analysis between them. Two kinds of test units were used in the experiment. The LoRa 868 MHz transmitter, which represented the unknown position, and the receivers that were used to estimate the position of the transmitter. The experiment gathered the RSSI values between the transmitter and receivers at different configurations and locations. The data collected from the experiment were analyzed using mathematical theory to answer the research question. The most accurate and precise value for each individual transmission parameter was established in the first testbed and used as the base data rates in the second testbed, to evaluate the best performing parameters simultaneously. The mean accuracy in testbed 1 varied from 180 cm to 388 cm, and the mean precision ranged between 0.432 dBm to 1.298 dBm. The mean accuracy in testbed 2 varied from 341 cm to 455 cm, and the mean precision ranged between 0.275 dBm to 1.495 dBm.  The experimental results indicate no connections between the data rate and precision. No correlation is found between the accuracy and the data rate. The standard deviation and absolute error fluctuate independently of the data rate and the transmitter position. In regard to the given results, the authors conclude that no linear relationship is found between the LoRa modulation configuration and the precision and accuracy of a position, in the selected environment. The experimental results show that LoRa could be used for indoor positioning applications where a rough estimation of a position is adequate since the mean accuracy is quite low for almost all tested modulation configurations. There could be applications where other factors, such as the energy consumption or communication range, are of more importance than accuracy. For those applications, LoRa could still be an adequate choice of technology.