Performance Analysis of Unauthorized UAV Detection Using RF Sensing under Non-Ideal Transceiver Conditions

While uncrewed aerial vehicles (UAVs) offer significant benefits in both civilian and military applications, their unauthorized use poses serious security threats. Their operation in restricted environments can compromise safety, security, and privacy, even though these vehicles are valued for their low cost, mobility, and ability to maintain line-of-sight (LoS) connections in non-terrestrial networks. Among the available countermeasures, radio frequency (RF)-based sensing is particularly attractive due to its affordability and effectiveness in both LoS and non-line-of-sight (NLoS) conditions. This work develops two detection frameworks for identifying malicious UAVs in the presence of non-ideal transceivers and provides a comprehensive performance analysis in terms of detection and false-alarm probabilities. The first framework employs the Neyman-Pearson (NP) detection approach, deriving analytical expressions for the probability of detection and the probability of false alarm while explicitly incorporating the impact of the non-ideal transceiver. The second framework introduces an optimal detector based on the log-likelihood ratio test (LRT), and derives a detection threshold that accounts for transceiver non-idealities. This LRT-based detector is shown to outperform the conventional NP threshold under realistic operating conditions. In the low-power regime, the LRT-based detector demonstrates noticeable performance improvements over the NP detector, highlighting the benefit of adaptive threshold design under practical channel and hardware conditions. Simulation results validate the analytical expressions and show that transceiver non-idealities degrade detection performance, particularly in LoS environments. For example, the probability of detection decreases by about 13.0% in suburban scenarios and 12.1% in urban scenarios, demonstrating the impact of transceiver non-idealities across different propagation environments. The findings highlight the need for impairment-aware design strategies for reliable UAV detection in real-world deployments.