IDS4IoT

Computational and Artificial Intelligence Solutions for Intrusion Detection in Internet of Things

The Internet of Things (IoT) is a fairly recent communication paradigm in which the objects of everyday life can communicate with one another and with the end users, becoming an internal part of the Internet.

There are plenty of definitions of the term ’Internet of Things’ (IoT) according to the global research and industrial community, which can be summarized as follows: “The Internet of Things is a network of physical objects (devices, vehicles, equipment, homes, buildings) that are connected to the internet through embedded devices and software, which allows these physical objects to collect, analyze, exchange data, and even make decisions on behalf of users”. There are various of technologies which are included in IoT embedded devices, such as wireless sensor networks (WSNs), intelligent sensing, remote sensing, Radio Frequency Identification (RFID), Near Field Communications (NFC), low energy wireless communications. The use of IPv6 technology enables these objects to connect and exchange data and enables them to inter-operate within the existing Internet infrastructure”.

The ease of communication between things through the Internet enables ease of access and interaction with a variety of devices, boosting the development of many applications that provide new services to individuals and organizations. Even though society can benefit from this publicly available burst of sensing data, new security challenges are created. Networks of this type inevitably attract people with malicious intent who aim in disrupting their functionality. A compromised object within the network may result in disrupting the network’s flow as it can transmit faulty data or even not transmit data at all.

An overarching goal of this project is to explore in a multidisciplinary way the prominent areas of Security, Computational/Artificial Intelligence and the Internet of Things s in the setting of a Smart City application.

The three research areas are interconnected in the creation of Intrusion Detection Systems. Intrusion Detection Systems (IDS) are found at the second line of security defense. A lightweight IDS anomaly detection mechanism for constrained sensor nodes has already been developed by us. Its main characteristic is the use of local agent information and Binary Logistic Regression (BLR) as the basis of detection of abnormal sensor behavior. BLR was trained for specific network layer attacks and it has an accuracy level between the range 96% and 100%.

The goal of this project is (a) to extend prior work into the Internet of Things field, (b) to implement additional anomaly detection mechanisms to detect attacks using Artificial intelligence and computational intelligence, and (c) examine all possible IoT network configurations in order to employ global and gateway agents for network data collection.