np-CECADA: Enhancing Ubiquitous Connectivity of LoRa Networks

Long Range Wide Area Networks (LoRaWAN) offer ubiquitous communications for The Internet of Things (IoT). However, there are many challenges in rolling out LoRaWAN - mainly scalability, energy efficiency, Packet Reception Ratio (PRR), and keeping the channel access as simple as unslotted ALOHA. To this end, we design non-persistent Capture Effect Channel Activity Detection Algorithm (np-CECADA), which is a novel, distributed protocol for the MAC layer of LoRaWAN. It utilizes Channel Activity Detection (CAD), which is a built-in imperfect mechanism for channel sensing and minimal feedback from the gateways. In np-CECADA each device independently adapts backoff times based on the traffic in its vicinity and the transmission power based on the heuristically inferred probability of capturing the channel. To achieve this, first, we carried out an extensive on-field evaluation to measure the effectiveness of CAD and capture effect in LoRa. Using them we designed np CECADA and developed $ns-3$ modules. Packet Reception Ratio of np-CECADA is $ 15.74\times$ and $ 5.13\times$ higher than vanilla LoRaWAN and p-CARMA, respectively. Channel utilization is $ 11.24\times$ higher compared to LMAC. Further, on a testbed of 30 LoRa devices np-CECADA outperforms LoRaWAN up to 5 times.