Deep Attention Recognition for Attack Identification in 5G UAV Scenarios: Novel Architecture and End-to-End Evaluation

Despite the robust security features inherent in the 5G framework, attackers will still discover ways to disrupt 5G unmanned aerial vehicle (UAV) operations and decrease UAV control communication performance in Air-to-Ground (A2G) links. Operating under the assumption that the 5G UAV communications infrastructure will never be entirely secure, we propose Deep Attention Recognition (DAtR) as a solution to identify attacks based on a small deep network embedded in authenticated UAVs. Our proposed solution uses two observable parameters: the Signal to Interference plus Noise Ratio (SINR) and the Received Signal Strength Indicator (RSSI) to recognize attacks under Line-of-Sight (LoS), Non-Line-of-Sight (NLoS), and a probabilistic combination of the two conditions. Several attackers are located in random positions in the tested scenarios, while their power varies between simulations. Moreover, terrestrial users are included in the network to impose additional complexity on attack detection. Additionally to the application and deep network architecture, our work innovates by mixing both observable parameters inside DAtR and adding two new pre-processing and post-processing techniques embedded in the deep network results to improve accuracy. We compare several performance parameters in our proposed Deep Network. For example, the impact of Long Short-Term-Memory (LSTM) and Attention layers in terms of their overall accuracy, the window size effect, and test the accuracy when only partial data is available in the training process. Finally, we benchmark our deep network with six widely used classifiers regarding classification accuracy. The eXtreme Gradient Boosting (XGB) outperforms all other algorithms in the deep network, for instance, the three top scoring algorithms: Random Forest (RF), CatBoost (CAT), and XGB obtain mean accuracy of 83.24 %, 85.60 %, and 86.33% in LoS conditions, respectively. When compared to XGB, our algorithm improves accuracy by more than 4% in the LoS condition (90.80% with Method 2) and by around 3% in the short-distance NLoS condition (83.07% with Method 1).