Design and Evaluation of a Security-Integrated Anomaly Detection Framework for IoT-Based Blood Bank Cold Chain Kits

This paper presents a security-integrated anomaly detection framework for IoT-enabled blood bank cold chain logistics. Blood supply integrity depends on maintaining strict environmental conditions during storage and transport, which are monitored using IoT sensors. These systems are vulnerable to both operational anomalies (e.g., temperature, humidity, geofence breaches) and cyberattacks that manipulate telemetry (e.g., spoofing, replay, suppression). Existing methods like Isolation Forests and risk fusion models detect statistical outliers or weight anomalies using Common Vulnerability Scoring System(CVSS) / Exploit Prediction Scoring System (EPSS) / Device Vulnerability Density (DVD) scores, but fail to capture malicious data manipulations and often raise false positives. We propose a risk-aware lightweight, layered anomaly detection framework that integrates per-feature temporal detectors, a cyberattack detector based on cross-sensor consistency, and a fusion module incorporating device vulnerability scores. Using a simulated IoT dataset, we demonstrate improved anomaly detection, reduced false positives through temporal persistence, and the ability to flag malicious manipulations. The proposed framework advances IoT cold-chain monitoring with a security-aware and explainable approach.