Self-Monitoring Obfuscated IoT Network
Ajay Biswas, Dipnarayan Das, Souvik Bhattacharyya · 2020
Internet of Things (IoT) has made automation cheaper and widespread, and future cities will be heavily dependent on it. Although IoT has made our life easier by providing hassle-free services and requiring negligible maintenance, it is vulnerable to hacking and physical abuse. In this paper, various protocols for secure communication between IoT and servers have been proposed. The communication takes place by using dual layers; one uses conventional communication techniques whereas the other uses a steganography channel for covert communication. These devices share their current status among each other, building a large AI network that can sense hacking and suspicious behaviour, and take necessary actions to keep the integrity of the system. The obfuscated behaviour of the network was demonstrated using various machine learning classifiers whose job was to correctly identify images that have a signature embedded within it.