Time Difference of Arrival (TDoA)-Based Localization in Mobile LoRaWAN Networks

Device localization remains a significant challenge in the Long-Range Wide Area Network (LoRaWAN) and is crucial for applications such as location tracking, emergency response, and real-time monitoring. A high number of localization techniques that exploit different information, such as Received Signal Strength (RSS) or packet arrival time, and based on different principles, such as Time of Arrival (ToA) or Angle of Arrival (AoA), have already been proposed. However, the performance achievable by these techniques degrades in case the node to localize is moving. In this case, would allowing one of the LoRaWAN gateways to move help improve the achievable performance? This paper proposes a Time Difference of Arrival (TDoA) localization based on the Non-linear Least Squares (NLS) technique and offers a performance evaluation comparing the proposed solution with other two TDoA solutions based on the extended Kalman Filter (EKF) and Particle Filter (PF). Different IoT device and LoRaWAN gateway mobility scenarios have been considered, how hyperbolic intersection geometry affects localization accuracy. The results indicate that NLS consistently outperforms EKF and PF in most considered scenarios, but its performance deteriorates under linear mobility due to poor gateway-node geometries.