Neural Network-based Cyber Threats Detection and Mitigation Framework in Aerial Communication Systems Underlying 6G

This study aims to enhance the security of communication between aeroplanes and ground control from potential risks. We introduce a novel framework, i.e., NeuroSkyWatch to protect aeroplanes from interference attacks like jamming and Distributed Denial of Service (DDoS). NeuroSkyWatch utilizes Bidirectional Long Short-Term Memory (Bi-LSTM), a neural network (NN), to detect and prevent threats proactively. Operating as an intelligent system, NeuroSkyWatch acts as a smart radar, continuously analyzing digital communication patterns in aerial systems using the Bi-LSTM model. Moreover, the Bi-LSTM’s ability to encompass past and future contextual information enables proactive threat detection. Additionally, NeuroSkyWatch leverages 6G technology to strengthen the security of airborne platforms in real time. This approach supports terabit-level data rates, ultra-low delay in the microsecond range, and extensive sensor/actuator data support. By combining NN algorithms with 6G networks, NeuroSkyWatch strengthens defences against evolving cyber threats in aerial environments. NeuroSkyWatch demonstrates its effectiveness across diverse scenarios, achieving a prediction accuracy of 95%, a communication delay of 245 milliseconds, and a validation loss of 0.31, surpassing existing approaches in security and performance.