Dynamic Trust Evaluation Framework for IoT

Abstract: The rapid expansion of the Internet of Things (IoT) has introduced significant security and trust management challenges due to the heterogeneous and dynamic nature of interconnected devices. Conventional trust evaluation approaches often rely on static parameters or recommendation-based mechanisms, making them vulnerable to malicious behaviors, false feedback, and evolving cyberattacks. This paper presents a dynamic trust evaluation framework that combines Quality of Service (QoS)-based metrics with entropy-driven weighting and reputation-based trust updating to provide adaptive trust assessment in IoT environments.

The proposed framework consists of five major stages: data collection, data processing, trust evaluation, trust validation, and trust updating. Network performance indicators including Packet Delivery Ratio (PDR), latency, and throughput are extracted from IoT communication traces generated through simulation environments. The collected metrics are normalized and weighted using an entropy-based method to determine their relative importance dynamically. A trust score is then computed for each device and validated against an adaptive threshold to classify nodes as trusted or untrusted. The proposed approach enhances trust assessment accuracy, reduces susceptibility to manipulation attacks, and supports adaptive decision-making in heterogeneous IoT networks. By integrating information-theoretic trust computation with dynamic reputation management, the framework provides a scalable foundation for secure and resilient IoT deployments.

Keywords: Dynamic Trust, Entropy, Packet Delivery Ratio, Quality of Service