Fog-assisted IoT-enabled scalable network infrastructure for wildfire surveillance

Abstract

Forest fires frequently termed as wildfires are fiercely destructive disasters causing enormous ecological and economic damage, as well as the loss of human lives. Global predictions for increased incidence and destructiveness of forest fires due to warming climate, drought conditions, urbanization and arson highlight the importance of an effective forest fire mitigation and management approach. Internet of Things (IoT) is well suited to ubiquitously assess the time-critical parameters for effective and reliable prediction of forest fires. This paper presents a novel Fog-assisted IoT-enabled framework for early prediction and forecasting of wildfires. The framework includes proposals for efficient energy utilization of the resource-constrained sensors responsible for wildfire monitoring by adapting the sampling rate of Wildfire Causing Attributes (WCAs) at Fog Layer. Moreover, the time enriched sampled data is further analyzed at Cloud Layer for predicting and forecasting the susceptibility of a forest block to wildfire outbreak. In addition, the forest area (in hectares) that could possibly be burnt in the event of wildfire outbreak is also predicted. Experimentation and performance analysis of the proposed system reveal that high values of accuracy, sensitivity, specificity, and precision averaging to 95.45%, 96.08%, 94.63%, and 95.64% respectively are registered for wildfire susceptibility prediction. Furthermore, Mean Absolute Error (MAE), Mean Square Error (MSE) and Root Mean Square Error (RMSE) values averaging to 0.25, 0.25 and 0.5 respectively are registered for wildfire susceptibility forecasting. Lastly, the efficacy of the proposed framework can also be derived from the real-time alert generation in the event of high wildfire susceptibility level.