Adaptive Data Rate Techniques for Energy Constrained Ad Hoc LoRa Networks

Long Range (LoRa) is an emerging low-power wide-area network technology. LoRa messages can be transmitted with a variety of parameters including transmit power, spreading factor, bandwidth, and error coding rates. While adaptive data rate (ADR) capabilities exist in the LoRa wide-area network (LoRaWAN) specification, this work is motivated by a cattle monitoring application where LoRaWAN is not feasible. In this scenario, the mobility of the animal changes the optimal parameter selections, which are the settings that transmit the data with the lowest energy consumption. This work analyzes ADR techniques to most efficiently find the optimal data rate for a firmware update, although the techniques are still valid for any large data exchange. It extends the ADR to use frequency shift keying (FSK) when there is enough signal strength since Semtech LoRa integrated circuits support FSK mode. The work uses dynamic acknowledgements and timeout values to improve the convergence time. The paper experimentally validates an analytical transmit time model and then describes three different methods for accomplishing the adaptive data rate. The methods are modeled analytically for the different convergence settings and two are demonstrated using the Microchip SAMR34 Explained boards.