A New Access Control Mechanism for IoT Based on Human Activity

The Internet of Things (IoT) is a rapidly growing field in technology due to its practical applications in daily life. It enables remote monitoring and control of electronic devices in a smart way, offering numerous benefits across various domains. The adoption of IoT devices and sensors is steadily increasing that leads to greater interaction among system components. However, IoT technology faces significant security challenges, as it is still in its early stages of development. In our research, we consider the access control challenge in IoT, where human activity is a significant component. We use machine learning algorithms and access control models on human activity data that are collected from various sensors to manage access to IoT devices. Extreme gradient boosting, decision tree, k-nearest neighbour, logistic regression, and random forest algorithms are trained and verified with ScientISST MOVE dataset. Analysis results show that the Random Forest algorithm outperforms other models of human activity from body sensor data. An attribute based access control model combined with a random forest model is more suitable to distinguish the types of data.