Federated Learning-Based Black Hole Prevention in the Internet of Things Environment

Abstract

The Internet of Things offers ubiquitous automation of things and makes human life easier. Sensors are deployed in the connected environment that sense the medium and actuate the control system without human intervention. However, the tiny connected devices are prone to severe security attacks. As the Internet of Things has become evident in everyday life, it is very important that we secure the system for efficient functioning. This paper proposes a secure federated learning-based protocol for mitigating BH attacks in the network. The experimental result proves that the intelligent network detects BH attacks and segregates the nodes to improve the efficiency of the network. The proposed techniques show improved accuracy in the presence of malicious nodes. The performance is also evaluated by varying the attack frequency time.