Aero-SecureNet: Transformer-powered Data Injection Attacks Detection Framework for 5G-enabled Aerospace IoT Networks

In recent years, the role of aerospace communication systems in today’s modern aerospace proven as a safe and efficient lifeline. However, aerospace communication systems remain vulnerable to sophisticated cyber threats as an open target for several attacks like Denial of service attacks, SQL injection attacks, etc. This may compromise the integrity of data and affect the communication of aircraft in the worst case. to handle this issue, we propose Aero-SecureNet framework, i.e., a deep learning-based framework that detects attack in Aerial and space networks (ASNs). The deep learning models employed are RNN, LSTM and Transformers, and perform a comparative analysis of these model. Optimization was carried out using several optimizers such as Adam, AdamW, and Nadam. Next, Aero-SecureNet is encompassed with 5G network to ensure ultra-Reliable Low-Latency Communication (uRLLC) to provide extremely reliable communication with latency as low as 1 millisecond. Aero-SecureNet analyzes the network traffic based on SQL injection attacks using 5G-uRLLC. Among the models, it was noticed that the Transformer performed better in the aspect of capturing contextual dependencies and processing sequential information. Aero-SecureNet identifies communication vulnerabilities to prevent the spread of erroneous data between stations and aircraft. Aero-SecureNet demonstrates an attack detection accuracy of 98.99% with respect to the existing approaches. Aero-SecureNet observed 32.97% of latency improvement in the proposed framework.