Cooperative Information Aggregation Based Distributed Sequential Estimation in Wireless Sensor Networks

The cooperative information aggregation (CIA) scheme for distributed estimation was proposed in our past research considering the realistic wireless sensor network (WSN) environments. In the existing CIA scheme, a multi-bit quantizer is applied at all nodes, however, each node only transmits one predetermined bit of its quantized observation to the fusion center (FC) to economize the energy expense and the communication bandwidth. In this work, the CIA based distributed sequential estimation (CIAS) scheme is proposed. In the proposed CIAS scheme, an instruction message is designed to be sent from the FC to the nodes sequentially to guide each node's transmitted bit selection. By sequentially polling the information bits from the nodes, the vote processes performed at the FC for the data fusion may be early terminated depending on the accumulative received information bits, and thereby numerous resource bits can be conserved in the estimation procedure. We design two algorithms for the CIAS scheme as well to pursue the promotion of energy conservation and estimation achievement. In the simulation results, considerable energy saving and estimation performance improvement in the proposed CIAS scheme are demonstrated, in comparison with the existing CIA scheme.