A novel intrusion detection system for internet of things devices and data

As we enter the new age of the Internet of Things (IoT) and wearable gadgets, sensors, and embedded devices are extensively used for data aggregation and its transmission. The extent of the data processed by IoT networks makes it vulnerable to outside attacks. Therefore, it is important to design an intrusion detection system (IDS) that ensures the security, integrity, and confidentiality of IoT networks and their data. State-of-the-art IDSs have poor detection capabilities and incur high communication and device overhead, which is not ideal for IoT applications requiring secured and real-time processing. This research presents a teaching-learning-based optimization enabled intrusion detection system (TLBO-IDS) which effectively protects IoT networks from intrusion attacks and also ensures low overhead at the same time. The proposed TLBO-IDS can detect analysis attacks, fuzzing attacks, shellcode attacks, worms, denial of service (Dos) attacks, exploits, and backdoor intrusion attacks. TLBO-IDS is extensively tested and its performance is compared with state-of-the-art algorithms. In particular, TLBO-IDS outperforms the bat algorithm and genetic algorithm (GA).