TDoA-based outdoor positioning in a public LoRa network Nico Podevijn 1 , Jens Trogh 1 , Abdulkadir Karaagac 2 , Jetmir Haxhibeqiri 2 , Jeroen Hoebeke 2 , Luc Martens 1 , Pieter Suanet 3 , Kim Hendrikse, David Plets 1 , Wout Joseph 1 1 Department of Information Technology, University of Ghent, IMEC, WAVES, Ghent, Belgium, nico.podevijn@ugent.be 2 Department of Information Technology, University of Ghent, IMEC, IDLab, Ghent, Belgium 3 Aucxis cvba, Stekene, Belgium Abstract—The performance of LoRa Geo-location for outdoor tracking purposes has been evaluated on a public LoRaWAN net- work. Time Difference of Arrival (TDOA) localization accuracy, probability and update frequency were evaluated for different trajectories (walking, cycling, driving) and LoRa spreading factors. A median accuracy of 200m was obtained and in 90% of the cases the error was less then 480m. Index Terms—LoRA, Geo-Location, Tracking. I. I NTRODUCTION Currently most track and trace solutions rely on GPS. Although providing accurate real time location updates, it has a disadvantage of consuming a fair amount of power and therefore a regular battery replacement is needed. As of today, public LoRaWAN networks are being deployed with Geo-location functionality which are capable of locating nodes using the Time difference of arrival (TDOA) technique [1]. The Geo-localization accuracy is worse than GPS but a main advantage is the fact nodes are able to operate for years without replacing the battery. Other advantages are the fact that a single technology is used for both communication and localization, making it a low-cost alternative. An example application for this alternative is tracking of goods between industrial warehouses. The novelties of this paper are: - First-time quantification of the TDOA Geo-location performance on a public deployed LoRa network for different scenarios (walking, cycling, driving). - Determination of the best spreading factor to use This paper is organized as follows. Section II describes related work in this new research area. In section III, the mea- surement setup, evaluated scenarios and estimation of TDOA performance are discussed. Results of TDOA performance and selection of best spreading factor to use are discussed in section IV. We summarize our observations and work in Section V. II. RELATED WORK In [4] a private LoRaWAN network with 4 gateways (each 2-3km apart) was deployed with TDOA capabilities and the geolocalisation was performed for stationary nodes. The re- ported mean accuracy was around 100m. The difference with our work lies in the fact that we investigate and improve the performance on a publicly deployed network with mobile Fig. 1. Walking (blue), cycling (green) and driving (red) Trajectories. The black pointers are the gateway locations nodes. To our knowledge, this has never been performed, mainly due to the fact LoRaWAN public networks with TDOA capabilities have only become recently public available [1]. III. MATERIALS AND METHODS A. Configuration Six LoRa nodes were configured and provisioned in the public LoRa network in the Netherlands. For each node ADR (Adaptive Data Rate) wad disabled to force a fixed spreading factor (SF7-SF12). Uplinked data size was set to 2 bytes (e.g. for sensor data and battery level status). Our implemented application server recorded the JSON messages from the network server which hold the location estimates (in latitude and longitude degrees) for each device. Since the time on air depends on the spreading factor used [5], the period between uplinks was set accordingly in order to respect the duty cycle of 1% in the 868MHz industrial, scientific and medical (ISM) band. Table I shows which transmission interval or period between transmissions was used for each node.