LSTM-Based Jamming Detection and Forecasting Model Using Transport and Application Layer Parameters in Wi-Fi Based IoT Systems

The Internet of Things (IoT) is increasingly used in every application field. Wi-Fi-based communication is the most preferred choice in wireless IoT systems due to its accessibility, high bandwidth, affordability, and versatility. However, security vulnerabilities in IoT devices pose considerable risks, especially in industrial and critical applications. This paper presents a novel technique to detect jamming in IoT, focusing on communication and environmental dynamics in industrial settings. To better examine the communication conditions and jamming in the industrial production environment, it initially focuses on collecting quality of service (QoS) and connection parameters during normal communication and jamming attacks in production lines equipped with wireless IoT devices with server-client architecture. Experimental findings demonstrate the strikingly profound impact of spot jamming on the upper layers of the protocol stack, revealing that jamming attacks on the physical layer can trigger cascading failures at the transport and application layers. This study presents an innovative stacked Long Short-Term Memory (LSTM)-based model for jamming detection that uses parameters from multiple layers. The proposed LSTM model demonstrates remarkable accuracy in predicting jamming attacks and exhibits robust performance in real-time IoT applications. The model is trained on a carefully curated dataset and tested on a real-time experimental testbed with various types of jamming attacks. The results yielded a jamming detection accuracy of 99.5% and a precision of 99.4%. This research sheds light on the impact of jamming attacks in IoT systems and presents an advanced detection model that shows promising results in mitigating such threats.