Neuro-Fuzzy Method for Predicting and Assessing the Level of Technological Hazard in Cyber-Physical Civil Protection Systems

Abstract

The paper proposes a neuro-fuzzy method for predicting and assessing technological hazard levels in cyber-physical civil protection systems, combining fuzzy inference with adaptive neural-network learning for processing heterogeneous sensor data and forming an integrated risk index. The proposed approach enables assessment of the current hazard level and prediction of its dynamics over a specified time horizon, supporting the transition from reactive monitoring to proactive detection of pre-accident and hazardous states. The feedback loop provides adaptive refinement of the neuro-fuzzy model using new data, which increases its applicability to various critical and industrial infrastructure facilities and supports informed decision-making in civil protection systems.