September 28, 2026
We are pleased to announce that our paper “SHIoT'26: A Specialized IoT Intrusion Detection Dataset for Smart Home Environments” has been accepted at the Nordic Conference on Secure IT Systems (NordSec) 2026.
The paper presents a new dataset for evaluating security solutions in realistic smart home IoT environments. The work contributes to the ongoing research on IoT security, intrusion detection, and machine learning-based anomaly detection.
We look forward to presenting our work at NordSec 2026 and contributing to discussions on the future of secure and resilient IoT systems.
Paper abstract:
The rapid integration of Internet of Things (IoT) devices into consumer spaces has significantly expanded the attack surface of modern smart homes, necessitating robust, data-driven security solutions. However, the development of effective Intrusion Detection Systems (IDS) is frequently bottlenecked by a lack of realistic, contemporary attack datasets.
To address this gap, the paper introduces a novel IoT network attack dataset captured in a realistic networking environment. The testbed incorporates 17 diverse real-world IoT devices, including cameras, environmental sensors, smart lighting, and power plugs, providing a heterogeneous smart home environment representative of real-world deployments.
The dataset contains 17 distinct network attacks covering three major categories:
DoS/DDoS attacks
Network scanning
Mirai-based botnet infections
In addition to releasing the complete dataset, the paper proposes a curated and optimized feature set designed to support efficient anomaly detection. To evaluate the dataset’s utility and establish baseline performance, the study conducts a comprehensive evaluation using five machine learning models and one statistical method.
The resulting dataset and feature set provide a valuable foundation for the development, evaluation, and deployment of next-generation security mechanisms for smart home IoT environments.