Speakers
Description
This study presents a hybrid approach that integrates probabilistic modeling and Failure Mode and Effects Analysis (FMEA) for the analysis and optimization of reliability and fault tolerance in AI-driven integrated security systems. With the increasing complexity of modern intelligent systems—incorporating video analytics, Internet of Digital Things (IoDT) sensors, autonomous drones, and distributed computing infrastructures—there is a growing need for comprehensive methods to evaluate system risk and resilience.
A mathematical model based on stochastic processes and Markov chains is developed, enabling quantitative evaluation of failure probabilities and transitions between key system states, including normal operation, degraded mode, and complete failure. In parallel, an FMEA analysis is applied to identify critical components, assess the consequences of potential failures, and calculate Risk Priority Numbers (RPN).
The proposed hybrid model integrates probabilistic assessments with expert-driven engineering analysis, resulting in improved accuracy in reliability evaluation. A system-level analysis is conducted across the main subsystems, including the sensing layer, AI processing modules, communication infrastructure, and robotic response units. The results indicate that the communication network and AI modules represent the most critical elements in terms of failure risk.
The proposed methodology can be applied in the design, optimization, and management of intelligent security systems, contributing to enhanced robustness, reliability, and operational efficiency under real-world conditions.