<p>Physical-layer authentication (PLA) in multi-user wireless sensor networks (WSNs) faces significant challenges in mixed-velocity vehicular environments, where the discriminability of physical-layer attributes varies substantially with user mobility. In particular, conventional single-attribute schemes cannot provide reliable authentication over the full velocity spectrum: channel state information (CSI)-based authentication degrades at high velocities due to rapid channel decorrelation, whereas Doppler shift (DS)-based authentication becomes ineffective at low velocities when an intelligent attacker closely matches the target motion. To address this issue, this paper proposes a velocity-adaptive dual-attribute authentication scheme that jointly exploits CSI and DS under intelligent tracking Sybil attacks. First, an intelligent attacker model is established to characterize the attack capabilities and physical limitations in mixed-velocity scenarios. Then, closed-form false alarm rate (FAR) and missed detection rate (MDR) expressions are derived for CSI- and DS-based schemes, revealing a fundamental velocity-security bottleneck and establishing the necessity of dual-attribute authentication. Based on this insight, two complementary decision strategies are designed over different velocity regions, where an AND-rule is adopted in the medium-velocity region and an OR-rule is used in the low- and high-velocity regions. The velocity partition and switching criteria are analytically determined from a closed-form performance metric. Closed-form FAR and MDR expressions are further derived for the proposed scheme. Simulation results validate the analysis and show that the proposed scheme achieves an authentication rate of 92.02%, consistently outperforming single-attribute methods across diverse mixed-velocity scenarios.</p>