Whitepapers

LoRaWAN ™ and LoRa Alliance ™ are marks used under license from the LoRa Alliance ™ www.lora-alliance.org GEOLOCATION WHITEPAPER THE LoRaWAN™ GEOLOCATION CAPABILITY This section provides a technical overview of LoRaWAN TDOA geolocation and error factors impacting accuracy. 4.1 Architecture A LoRaWAN end-device can be located if uplink transmissions from the device are received by three or more gateways. These uplink transmissions need not be specific transmissions for geolocation; they may be typical LoRaWAN application data frames. Several gateways simultaneously receive the same uplink message, and the end-device location is determined using multilateration techniques. The multilateration process is shown in Figure 4-1. There is no additional hardware required on the end-device beyond its LoRaWAN interface. Gateways do require accurate time synchronization; this is currently achieved with GPS at the gateways (or any means available to synchronize gateway clocks to within a few tens of nanoseconds). Each received uplink frame is accurately time-stamped by the gateway. This time stamp is forwarded to the network server as part of a frame's metadata, which also includes signal level, signal-to-noise ratio and frequency error. The network server sorts multiple receptions of the same frame, groups all the metadata including the timestamps for this frame, and requests a geolocation computation from the geolocation solver. The elementary geolocation solver function is to compute, for a given frame, the difference in time of reception seen by pairs of gateways. This time difference measures proximity of the end-device to one gateway of the pair compared with the other. When the TDOA is known for a pair of gateways, the end-device can be placed on a hyperbola. With (i) several such time differences, the end-device can be placed on several hyperbolae. The end-device is positioned at the intersection of these hyperbolae. The accuracy of the position fix depends on several factors: • Propagation environment and multipath • Gateway deployment geometry and density • Position determination algorithm used by the geolocation solver • Quality of gateway's time synchronization • End-device dynamics and configuration 4.2 Impact of Propagation Errors In a multipath-free environment, LoRaWAN geolocation performance is limited by the gateway's clock accuracy. Conductive geolocation testing typically achieves better than 3m accuracy with signal levels 25dB above sensitivity. At sensitivity level, noise degrades performance to 60m. Within these constraints, accuracy rarely depends on the received signal level. In the presence of multipath, given the system bandwidth limitation of 125 KHz, signal paths are often indistinguishable. Only the average channel delay can be estimated. In some cases the direct signal path is not present, introducing a delay offset into the frame timestamps, as only reflection paths are seen. Figure 4-2 shows the statistics of timestamp errors for different propagation scenarios. These are measurements taken from mobile vehicle testing with timestamp errors estimated using GPS. The urban case has fewer data points than the others (2,000 vs. 10,000), which explains the worse-looking curve. The average distance from vehicle to gateway is 1.5km. Figure 4-1: Geolocation Architecture 4

• Cover