AVTONOM KO‘P QATLAMLI IoT TIZIMLARIDA ENERGIYA HOLATI VA NOSOZLIK XAVFIGA MOS O‘Z-O‘ZINI

Abstract

An energy–risk adaptive self-diagnosis method is proposed for
autonomous multilayer Internet of Things systems. It fuses sensor-, edge-, and
communication-layer evidence into a risk score and switches monitoring among lowpower,
balanced, and intensive modes according to anomaly risk and battery state.
Across 100 Monte Carlo runs of 30 simulated days, the method reduced energy
consumption by 56,2% relative to a fixed balanced policy while retaining a
comparable fault-episode recall (0,697 versus 0,709). The reported values are
reproducible simulation results and require subsequent hardware validation.