Edge Based Decision Making In Disaster Response Systems
Jonas Wagner, Mehdi Roopaei · 2020 10th Annual Computing and Communication Workshop and Conference (CCWC) · 2020
Disaster response systems utilizing device analytics can help rescue teams to identify needs, prioritize responses, and increase response efficiency. Machine Learning (ML) refers to methods that can processes vast amounts of data observed during crises and can also predict future events, such as the potential aftershocks of an earthquake or additional flooding. Critical analytics require time-sensitive data processing that make optimal use of distributed edge computing to ensure realtime computation. This paper discusses the current progress in developing edge analytic decision-making systems for aerial vision processing. The system will use AI-driven techniques to analyze aerial footage and recreate a real-time model of the environment. Most importantly, the system will be able to detect and identify individuals in distress and immediately alert disaster responders. In the proposed framework, disaster scenarios are created in the virtual domain and environmental information collected at the edge is used to develop and test decision making systems that can generate an in-depth model of the entire environment for disaster responders.