Universitas Dian Nuswantoro conducted research on an intrusion detection system for Internet of Things infrastructure using an ensemble learning approach and the synthetic minority over-sampling technique known as SMOTE. The activity involved UDINUS researchers Kurniawan, Aldhi Ari; Santoso, Heru Agus; Soeleman, M. Arief; and Fanani, Ahmad Zainul as participants in the research.
The research addressed the increasing use of IoT-based systems in the context of the industrial revolution 4.0. The source describes IoT use in health, manufacturing, urban planning, housing, and automotive fields. It identifies security considerations related to data integrity, privacy, and service stability. Through intrusion detection, activities on an IoT system can be analyzed to determine whether suspicious activities may interfere with or threaten IoT services.
The activity responded to the high false alarm rate associated with intrusion-detection approaches reported in several previous studies. UDINUS proposed the use of machine learning, specifically ensemble learning combined with SMOTE, as an approach for detecting intrusions in IoT systems. The research examined whether the proposed approach could provide better performance than base learning and approaches used in previous studies as intrusion-detection methods.
The reported results indicate that the proposed approach detected intrusions and classified them into five types: normal intrusion, probe, dos, r2l, and u2r. The evaluation reported an accuracy of 97.02%, a detection rate of 97%, and a false alarm rate of 0.16%. These results were reported as improvements in intrusion-detection performance compared with base learning and approaches from previous studies. However, the processing-time performance had not shown satisfying results.
The activity is related to innovation and infrastructure through machine-learning-based security analysis for IoT infrastructure. The source also identifies urban planning as one field in which IoT-based systems are used. The activity is institutionally framed as UDINUS research addressing intrusion detection and security considerations in IoT systems.