In today's world of technology, there are many types of IoT devices in use, with the number of devices increasing exponentially over the next few years. However, many of these devices generally lack sufficient protection and are prone
to compromise and forgery. An important step in improving IoT security is to uniquely identify the IoT devices in a network accurately. Identifiers such as IP addresses, MAC addresses, IMEI numbers, and more can be easily forged
by a malicious actor.
In this project, we aim to extract a more robust fingerprint by observing behavioral patterns of IoT devices in a network that can’t be forged easily throughout a fixed duration. The fingerprints will then be used to
train a machine learning model for automatic classification and accurate prediction of these devices, which opens the door to anomaly detection implementation.