Mitigating SIoT attacks in Smart Medical Systems: A Machine Learning based Approach

Abstract

The idea of Social Internet of Things (SIoT) is materializing as we find the opportunities and the need for blending Internet of Things (IoT) with concepts of social networking, where human control is relinquished for independent devices to act as autonomous systems. These systems allow “objects” to interact with other objects in the network devoid of any external supervision which demands that trustworthiness of devices be considered for quality communication. In this paper, the authors focus on various trust related attacks and present a trust management model based on machine learning for determining trust values of objects in a smart medical network (SMN). In the proposed SMN, the objects are capable of using the model as and when required. Principal Component Analysis (PCA) is used to determine the significance of metrics for direct trust computation. The proposed trust management model successfully mitigates seven SIoT attacks as demonstrated through the experiments.