文献类型：期刊文献

英文题名：Secure Distributed State Estimation for Connected Vehicles under Non-Uniform Sampling

作者：Wang, Zifeng[1]; Yang, Wen[1]; Ding, Wenjie[1]; Qi, Nuofan[1]

机构：[1] East China University of Science and Technology, Key Laboratory of Smart Manufacturing in Energy Chemical Process, Shanghai, 200237, China

年份：2026

外文期刊名：IEEE Transactions on Vehicular Technology

收录：EI(收录号：20262721032760)

语种：英文

外文关键词：Communication channels (information theory) - Covariance matrix - Distributed computer systems - Errors - Frequency estimation - Gaussian distribution - Network protocols - State estimation - Stochastic systems - Structural frames - Vehicle to vehicle communications - Vehicles

摘要：Distributed state estimation in connected vehicles faces significant challenges arising from the non-uniform sampling inherent in Vehicle-to-Everything (V2X) communications and the threat of False Data Injection (FDI) attacks. Addressing the data irregularity caused by channel congestion and stochastic packet loss, this paper proposes a distributed estimator based on a lifting technique. Unlike interpolation-based approximations that degrade high-frequency dynamic information, this method aggregates asynchronous onboard measurements and V2X messages within an estimation frame to achieve superior fusion accuracy. Theoretical analysis establishes the monotonicity of the estimation error, showing that accuracy improves as more measurements are collected, and proves that the error converges within fixed upper and lower bounds. Furthermore, an efficient offline precomputation strategy reduces the high computational burden. In parallel with estimation, a Gaussian Mixture Model (GMM) detector is introduced, which is decoupled by operating solely on exchanged state estimates without requiring additional communication or access to internal covariance matrices. This design allows the detector to verify neighbor state estimates as they arrive, rather than waiting to receive all information after a measurement frame. Finally, numerical simulations and physical experiments verify that the proposed framework achieves superior accuracy and robust resilience against attacks. ? 1967-2012 IEEE.