Detection and Classification of Interference affecting LoRaWAN communications in Railway environment

The French national railway company (SNCF) is deploying Internet of Things (IoT) technologies using the LoRaWAN communication protocol to centralize and transmit the data measured by the onboard sensors to the railway maintenance centers. They recently developed a communication interface connected to the different sensors called MARTI. A gateway called MELI then switches data from the LoRa protocol to the 4G network to centralize them to SNCF’s IoT platform. However, the reception of the gateway MELI can be affected by the transient electromagnetic interference occurring with the catenary-pantograph contact losses. Moreover, in a security context, these communications can also be intentionally disturbed by the use of jammers. This work aims to detect the presence of this intentional and non-intentional interference and to distinguish them. This should allow sending the LoRa signal at instants without interference to guaranty the good reception of the LoRa communications by the gateway. We performed experiments in the laboratory to analyze the performance of a Support Vector Machine classification (SVMc) approach to detect and separate such interference.