Privacy-Preserving Distributed Beamformer Design Techniques for Correlated Parameter Estimation
Privacy-Preserving Distributed Beamformer Design Techniques for Correlated Parameter Estimation
复制标题
DOI：
10.1109/jsen.2023.3310658
复制
发表时间：
2023-11
影响因子：
4.3
通讯作者：
M. Ahmed;Kunwar Pritiraj Rajput;Naveen K. D. Venkategowda;A. Jagannatham;L. Hanzo
中科院分区：
文献类型：
--
作者：
M. Ahmed;Kunwar Pritiraj Rajput;Naveen K. D. Venkategowda;A. Jagannatham;L. Hanzo
针对基于正交频分复用（OFDM）的无线传感器网络（WSN）中的时间相关矢量参数估计，提出了隐私保护分布式波束形成设计。利用固有的时间相关性的参数向量的率失真理论为基础的比特分配框架用于传感器测量的最佳量化。提出的分布式波束形成设计是通过融合的双重共识交替方向乘法器（DC-ADMM）技术与相关的隐私保护框架。这使得每个传感器节点（SN）可以以分布式方式设计其发送预编码器，这最小化了重要信息对恶意窃听者（EV）节点的敏感性，同时避免了将状态信息发送到融合中心（FC）的集中式方法所需的显著通信开销。贝叶斯克拉美-拉奥界（BCRB）的基准建议的发射波束形成器和接收机组合器的设计的估计性能，而我们的仿真结果说明了性能，并明确展示了隐私和估计性能之间的权衡。
Privacy-preserving distributed beamforming designs are conceived for temporally correlated vector parameter estimation in an orthogonal frequency division multiplexing (OFDM)-based wireless sensor network (WSN). The temporal correlation inherent in the parameter vector is exploited by the rate distortion theory-based bit allocation framework used for the optimal quantization of the sensor measurements. The proposed distributed beamforming designs are derived via fusion of the dual consensus alternating direction method of multiplier (DC-ADMM) technique with a pertinent privacy-preserving framework. This makes it possible for each sensor node (SN) to design its transmit precoders in a distributed fashion, which minimizes the susceptibility of vital information to malicious eavesdropper (Ev) nodes, while simultaneously avoiding the significant communication overhead required by a centralized approach for the transmission of the state information to the fusion center (FC). The Bayesian Cramer–Rao bound (BCRB) is derived for benchmarking the estimation performance of the proposed transmit beamformer and receiver combiner designs, while our simulation results illustrate the performance and explicitly demonstrate the trade-off between the privacy and estimation performance.