Multiple Botnet and Keylogger Attack Detection Using CNN in IoT Networks
S.L Ananthu Suresh, Anu Susan Philip · 2022
Modern electronic devices are developed to be wireless and smart. These devices can communicate with each other and their host servers all day. Constant connectivity helps the user to have a real time knowledge about his/her environment. These devices are called IoT devices. With advancement in technology, cyber threats have also evolved. Latest malwares target IoT networks to steal user information for illegal activities. We are proposing a detection technique to identify most of the modern IoT network threats using CNN and machine learning. Then we are providing an evaluation on its performance by calculating Precision, Recall, Accuracy and F1score. Botnet and Keylogger attacks are considered for this project. Botnet attacks find vulnerabilities in an IoT device and then take control of its operations. Keylogger attacks target the user privacy and confidential credentials like bank records, passwords, etc. The model was able to produce an accuracy of 90%. Finally, we provide a hardware demonstration to show how the IoT devices can detect intrusion attempt when under a botnet attack using Raspberry Pi and Node MCU (ESP8266).