A LoRa-Based Linear Sensor Network for Location Data in Underground Mining

Abstract

In this paper, we describe a LoRa (from “Long Range”)-based, linear sensor network we have developed for transmitting location information of personnel and equipment in an underground mine. The system is intended to be used during emergencies when existing communications infrastructure has failed. Linear networks comprise a sequence of relays that forward data to a common destination, the headend. Relays forward location information transmitted from tags carried by personnel or equipment. Relays will usually be put in place as investigators or rescuers enter the mine. LoRa is used both by the tags to communicate to the relays and by the relays to forward messages to the headend. We have implemented and tested this system, and have carried out simulations and analyses to determine its scalability, reliability and fairness. The need for robustness and reliability has led us to use flooding rather than unicast communication. We also use message sequence numbers and time-to-live fields to prevent broadcast storms. Contention is managed using a simplified Carrier Sense Multiple Access (CSMA) scheme. We also address fairness. When the network is under load messages may be dropped by relays making messages generated more hops from the headend more likely to be dropped than messages nearer the headend. We explore the relationship between unfairness, traffic load and number of relays. We also observe that a network of larger numbers of lightly loaded relays performs more effectively than smaller numbers of heavily loaded relays.