One-bit Parameter Estimation Algorithm In Multiplicative Noise Environment

Wireless sensor networks(WSNs)usually consist of a large number of tiny sensor nodes with integrated sensing,signal processing,and communication capabilities.Each node can collaboratively collect,analyze and transmit the physical information of interest and have a wide range of applications in target tracking,target positioning,environmental monitoring and other fields.The application problems based on WSNs can usually come down to the problem of parameter estimation based on WSNs.Normally,each sensor node sends the sensing information to the fusion center.The fusion center uses the received data to design the corresponding algorithm,analyze and process these data,so as to obtain estimation of the unknown parameter.However,the sensing resolution,computational and communication capabilities,and storage capacity of each sensor node are limited.This presents a greater challenge to the design of parameter estimation algorithms based on WSNs.In order to avoid a large bit of information transmission,and hence resulting in more energy consumption and larger communication load of the node,the paper considers compressing the sampled signal value of each sensor node to one-bit information,and then transmitting it to the fusion center for centralized processing.In addition,in order to better adapt to the actual environment,the paper considers that the sampled signals of the node are corrupted by both additive noise and multiplicative noise.Combining these problems,the paper proposes a one-bit maximum likelihood estimation algorithm based on expectation maximization(EM),which can obtain the similar estimation accuracy as the maximum likelihood estimation(MLE)algorithm based on analog measurement values(infinite bit accuracy),and the mean square error of the estimated values is close to the Cramér-Rao lower bound.Note that the adaptive signal processing has lower storage requirements and is suitable for online parameter estimation,especially for slowly time-varying parameters,this paper proposes an recursive least square(RLS)algorithm for one-bit measurement in a multiplicative noise environment.On the other hand,for large-scale wireless sensor networks,the number of sensor nodes is large and widely distributed,and the traditional centralized wireless network has more disadvantages.In contrast,distributed sensor networks have better robustness.Therefore,this paper further studies the distributed parameter estimation problem of sensor networks based on one-bit sampling signals in multiplicative noise environment,and proposes distributed parameter estimation algorithms based on threshold optimization and expectation maximization,respectively.The paper performed a large number of simulation experiments on the proposed algorithm and also compared it with other existing algorithms.These experiments and analysis results verify the effectiveness of the algorithm in this paper.