Monitoring and Smart Decision Architecture for DRONE-FOG Integrated Environment

• DOI: 10.5753/sbcup.2021.16008
• Publication Date: 2021-07-07T16:36:24Z
• Authors: Wendel Serra; Warley Junior; Isaac Barros; Hugo Kuribayashi; João Carmona
• Keywords: Drone; Cloudlet; Fog Computing; Perceptron; Computation offloading; Multilayer perceptron

Due to the limited computing resources of drones, it is difficult handle computation-intensive tasks locally, hence, fog-based computation offloading has been widely adopted. The effectiveness an operation, however, determined by its ability infer where execution code/data represents less computational effort for drone, so that, deciding offload correctly, device benefits. Thus, this paper proposes MonDroneFog, a novel architecture that supports image offloading, as well monitoring and storing performance metrics related wireless network, cloudlet. It takes advantage main machine-learning algorithms provide decisions with high levels accuracy, F1, G-mean. We evaluate classification under our database results show Multi-Layer Perceptron (MLP) Logistic Regression classifiers achieve 99.64% 99.20% respectively. Under these conditions, MonDrone-Fog works in dense forests when weather conditions are favorable can be useful support system SAR missions providing shorter runtime operations.