Magnifier: Leveraging the Fine-Grained Hardware Information and Their Temporal Patterns for Concurrent LoRa Decoding

LoRa Wide Area Network (LoRaWAN) is famous for its low power consumption and wide coverage area. A critical issue that constrains the scalability of LoRaWAN is how to support concurrent packet reception efficiently. To this end, our work first carefully investigates all the hardware imperfections between a LoRa transceiver pair and studies their impacts on the received signal's temporal patterns with fine granularity. Such imperfections and the associated temporal patterns are intrinsic and unique for any transceiver pair in a wireless system. Therefore, they provide a brand-new and highly influential perspective for identifying different packets. Motivated by this, we propose Magnifier, which integrates the two factors mentioned above for LoRa concurrent decoding. Though some prior work leverage partial hardware imperfections, they fail to systematically study and exploit the temporal patterns caused by the imperfections for concurrent packet reception. We also evaluate the performance of Magnifier with 40 transmitters (commodity LoRa chips) and one receiver (USRP B210). Both critical complexity analysis and extensive experiments show that with the same hardware settings and the same scale of complexity, Magnifier can support 3× to 8× throughput of other existing schemes.