基于分布式无线传感器最优比特分配

Optimal bit allocation based on distributed wireless sensor networks

当无线传感器网络的总比特数固定时,给出表现优良的无偏量化器,探究了静态参数的量化估计与整个网络的最优比特数分配,与提前假定传感器量化比特数做法大不相同.首先给出一种新型的无偏量化器,进而得到理想的线性量化观测估计器及其均方误差上限.而后通过理论分析发现当整个网络总带宽固定时,最优比特的分配取决于传感器的信噪比,而且总量化级数与信噪比的比值在一定程度上决定了整个网络中传感器的激活率.最后,模拟仿真验证了所提方法有效提高了最优比特分配估计器的性能.

When the total number of bits in the wireless sensor network is fixed,this article gives an excellent unbiased quantizer,and explores the quantization estimation of static parameters and the optimal bit allocation of the entire network,which is very different from presuming the sensor quantization bit number in advance.This article firstly presents a new type of unbiased quantizer,and then obtains the ideal linear quantized observation estimator and the upper bound of mean square error.Then through theoretical analysis,it is found that when the total network bandwidth is fixed,the optimal bit allocation depends on the signal-noise ratio of the sensor,and the ratio of the total quantization level to the signal-noise ratio determines the activation of the sensor in the entire network to a certain extent.Finally,simulations verify that the method in this study effectively improves the performance of the optimal bit allocation estimator.