Experimental vs. Simulation Analysis of LoRa for Vehicular Communications Fernando M. Ortiz a , Thales T. de Almeida a,b , Ana E. Ferreira a,c , Lu´ ıs Henrique M. K. Costa a a GTA/PEE/COPPE – Universidade Federal do Rio de Janeiro b CEFET/MG - Centro Federal de Educa¸ c˜aoTecnol´ ogica de Minas Gerais, Brazil c CEFET/RJ - Centro Federal de Educa¸c˜ ao Tecnol´ ogica de Rio de Janeiro, Brazil Abstract LoRa, a Low Power Wide Area Network (LPWAN) technology, demands re- duced network infrastructure to cover a large area, with low power consumption. In the context of vehicular communications, LoRa has the potential to support monitoring and cooperative navigation applications, where small amounts of data can be transmitted asynchronously through the network. Nevertheless, the literature lacks evaluations of LoRa in the context of vehicular communi- cations. Thus, in this paper we analyze the performance of this technology operating in an urban mobility environment, using LoRa terminal devices em- barked on vehicles and a LoRa receiver unit, acting as infrastructure. Moreover, we evaluate the equivalence between experimental and simulated results, ob- tained from the communication link between the LoRa module inside a vehicle and a LoRa receiver comparing to those produced by NS-3 simulations. Three metrics of interest are evaluated: Packet Delivery Ratio (PDR), Packet Inter- Reception (PIR) time, and Received Signal Strength Indicator (RSSI). Field experiments are performed at the campus of the Federal University of Rio de Janeiro (UFRJ), Brazil. Results revealed that all the metrics evaluated in the simulated experiments are consistent with the results of the real experiments, however, we concluded that the model can be improved by looking for a stronger * Corresponding author Email addresses: fmolano@gta.ufrj.br (Fernando M. Ortiz), almeida@gta.ufrj.br (Thales T. de Almeida), ferreira@gta.ufrj.br (Ana E. Ferreira), luish@gta.ufrj.br (Lu´ ıs Henrique M. K. Costa) Preprint submitted to Computer Communications June 17, 2020