Secure Vehicular Communications via Noise-Aided FH: Robust Waveform Design and Analysis

Traditional noise-aided frequency hopping (FH) architectures can secure autonomous driving against jamming and eavesdropping. However, their reliance on high-quality noise suppression incurs substantial computational complexity, clumsy noise storage and updating, and vulnerability to engineering impairments, thereby degrading the communication quality and straining computing resources for autonomous driving. This study proposes a novel noise-aided FH architecture with lower complexity and higher resistance to engineering impairments. Besides, the system secrecy with multiple eavesdroppers is assessed via the secrecy rate metric, and the average secrecy enhancement of the proposed architecture is 1.4 bps/Hz compared to the traditional architecture when the time error is 0.1 symbols. In addition, assuming that the channel state error (CSE) follows a zero-mean complex Gaussian distribution, the expectation and cumulative distribution function of the secrecy rate are derived, and the secrecy outage probability (SOP) is investigated to evaluate the secrecy reliability. Finally, the outage-based power allocation is investigated for two targets: (a) Minimizing SOP under a minimum data rate threshold, which reflects the quality difference between the legitimate and eavesdropping channels. The SOP of the proposed power scheme is 5.5% lower than that of the power scheme ignoring CSE, and 8.1% lower than that of the equal power scheme. (b) Maximizing the data rate under an SOP threshold, which reflects the secure transmission efficiency. The maximum data rate of the proposed power scheme is 0.15 bps/Hz higher than that of the power scheme ignoring CSE, and 0.25 bps/Hz higher than that of the equal power scheme.