Random Forest Algorithm based Device Authentication in IoT

Distributed devices and data management systems are major components of IoT network. IoT security and privacy are of the utmost significance because there are already billions of IoT devices, applications, and services in use and more are going online every day. Authentication, confidentiality, integrity and attacks are some important security and privacy issues in IoT systems. Authentication is a crucial security issue in order to secure the Internet’s devices and data. The process of identifying users or devices is known as authentication. In this paper, a non-linear machine learning algorithm is used for device authentication in the network. The proposed scheme adopts Belief-Desire-Intention (BDI) server agency with the Random Forest (RF) algorithm, using the context information of the IoT objects for authentication. The proposed authentication scheme operates as follows: Authentication request from IoT device to authentication server agency, generation of session key, upon request from an IoT device, validation of the IoT device using a username/password and session key. The authentication server calls the BDI model, which uses context based information of device to make authentication decisions using a random forest (classification and regression) algorithm. Simulation results show that the performance of the proposed intelligent authentication approach is better in terms of authentication accuracy, computational delay etc.