On Automated Trust Computation in IoT with Multiple Attributes and Subjective Logic

Abstract

Developing automated trust mechanisms has become crucial for overcoming perceptions of uncertainty and risk by people using IoT services. Things are increasingly communicating with each other and trust in the data they deliver depends on several factors such as the links they use to communicate and the environment. This points to a need for a trust management method for "things" that considers the communication among them, environmental and security-related factors, and the net-work topology but without human intervention. To address these challenges, we propose a trust management framework that automatically computes the trust of "things". We use Multi-Attribute Decision Making (MADM) and Evidence-Based Subjective Logic (EBSL) in a trust network of "things" to take into account the uncertainty in trust values. We propose new normalization for non-monotonic attributes in MADM. We present an algorithm for automatic trust computation and evaluate its effectiveness using synthetic data and sampling from real datasets.