Securing IoT Using Lightweight TCN for Edge Deployment on Raspberry Pi 4

The Internet of Things (IoT) is a network of tiny sensing devices that facilitate precision monitoring, automation, and intelligent decision-making in a connected digital environment. While IoT enables high connectivity, it also presents various issues relating to compliance, operational resilience, coexistence, and cybersecurity. These factors significantly affect the robustness, reliability, and resilience of IoT devices. To mitigate emerging cybersecurity issues, an efficient and lightweight security solution is essential for resource-constrained edge devices. This research addresses security challenges by leveraging Healthcare-IoT (H-IoT) Distributed Denial-of-Service (DDoS) datasets to deploy a lightweight TCN model on a standalone IoT device, the Raspberry Pi 4. It enables efficient edge deployment for detecting critical cyber threats, particularly DDoS attacks. The TCN model is converted into TensorFlow Lite (TFLite) format and optimized through quantization, reducing model size and computational overhead. The deployment achieves 99.95% accuracy with an average inference latency of 0.19 ms on the MQTT-based dataset, and 99.94% accuracy with 0.27 ms on the UDP-based dataset. The average power consumption on the Raspberry Pi 4, measured using a physical USB power meter, is 4.22 W and 4.64 W for the two datasets. These results demonstrate the feasibility of high-accuracy, resource-efficient DDoS attack detection on low-power edge devices for securing IoT, Industrial IoT (IIoT), and H-IoT systems in real-world environments.