Low-Cost Car Park Localization Using RSSI in Supervised LoRa Mesh Networks

In this paper, we present an RSSI-based localization solution that merely relies on the LoRa technology without requiring any anchors. It is primarily designed for large used car dealerships. Cars to be tracked are fitted with battery- powered tags that send each other plain LoRa messages and then report over LoRaWAN the observed RSSI to the server. The server estimates tag coordinates using multi-dimensional scaling and a novel corrective transformation based on tag clustering. We designed a low-cost tag for RSSI measurement and an energy efficient solution for managing a mesh of 1000 tags on one site using time division multiplexing. The tags are fully controlled and synchronized by the server. To stay synchronized for several hours until a next downlink the tags estimate and compensate the real-time clock error. Large-scale tests are yet to be performed, but lab-based and small-scale tests in real conditions show a maximal error of 8 meters and a battery lifetime of 5 years.