Cooperative Information Aggregation for Distributed Estimation in Wireless Sensor Networks

The performance of distributed estimation in wireless sensor networks (WSNs) is affected by observation noise, communication errors and the noise caused by the inevitable quantization process. Reducing the number of transmitted bits per node can minimize the energy consumption and communication load, thereby prolonging the network lifetime. However, reducing the number of transmitted bits has a significant impact on the estimation performance. In this work, we propose cooperative information aggregation (CIA) approaches for distributed estimation that consider the above-mentioned issues jointly. Under the proposed schemes, a multi-bit quantizer is used to generate the local observation codeword at each node. Then, each node only transmits one predesigned bit of the derived codeword to the fusion center (FC). At the FC, the information bits transmitted cooperatively by all nodes are aggregated and fused to obtain the estimation result. We design two estimators: an aggregation hard decision estimator (AHDE) and an aggregation maximum-likelihood estimator (AMLE). Predetermining the transmitted bit at each node is formulated as a constrained optimization problem for resource allocation and the corresponding suboptimal resource allocation policies are proposed. The simulation results demonstrate that the proposed schemes can overcome the impact of observation noise and communication errors effectively.