Intelligent Data Forwarding Scheme for LoRa based Fog Enabled Smart Agriculture

Nowadays, Internet of Things (IoT) technologies are very useful to transform traditional farming practices into smart farming across the globe. Since most of the agricultural land is located in remote places, there is a need for new technologies for making them smart. For smart farming, low-cost, and resource-constraint sensors are placed across the agricultural field. To provide continuous connectivity for data communication, there must be a suitable and reliable transmission over long distances and, at the same time, should consume less energy. In particular, Long Range (LoRa), a relatively new communication technology, uses long waves to work over long distances. This is extremely useful in agriculture, where the communicating areas are broad fields of crops and Internet connectivity is low and/or intermittent. So, to reduce data transmission between layers (sensing, fog and cloud), a machine learning based data forwarding architecture is proposed for the three tier architecture, namely, Regressive Prediction Data Forwarding Model (RPDM). A suitable and lightweight machine learning model is deployed on each layer which will predict the next coming sensing data. So, the upcoming data will not be forwarded from the sensing layer to the fog layer if they match with the predicted one. Also, in the case of Internet failure, the sensing layer made capable of performing the actuation based on current data value. We validated the performance of RPDM on a real testbed in terms of the amount of energy consumed and data transmitted.