LSTM-Based Jamming Diagnosis and Prediction Model Utilizing Transport in Wi-Fi-Enabled IoT Systems

Every industry is seeing a rise in the use of the Internet of Things (IoT). Because of its widespread availability, high bandwidth, low cost, and adaptability, Wi-Fi-based communication is favored by wireless IoT devices. However, serious dangers exist due to security flaws in IoT devices, particularly in mission-critical and industrial settings. With an emphasis on industrial communication and environmental dynamics, this article introduces a new method for detecting jamming in the Internet of Things (IoT). Using production lines that include wireless IoT devices with a server-client architecture, it first focuses on collecting connection parameters and quality of service (QoS) during normal communication and jamming attacks so that it can more effectively investigate the communication conditions along with jamming in the industrial production environment. Results from experiments show that spot jamming has a devastating effect on higher-level protocol layers, showing that physical-layer jamming assaults may cause transport and application-layer failures to cascade. This research introduces a novel multi-layer jamming detection model based on stacked Long Short-Term Memory (LSTM). With its strong performance in real-time applications for IoT and impressive accuracy in anticipating jamming attempts, the suggested LSTM model is a clear winner. The model undergoes testing on a real-time experimental testbed that simulates different sorts of jamming assaults after being trained on a dataset that has been meticulously selected. A 99.5% accuracy rate for jamming detection and a 99.4% precision rate were produced by the findings. This study provides new insight into the effects of jamming assaults on Internet of Things (IoT) devices and introduces a sophisticated detection model that has the potential to reduce the severity of these attacks.