Does More Transmitting Sensors Always Mean Better Decision Fusion in Censoring Sensor Networks with an Unknown Size?

Abstract

This paper examines the impact of sensor censoring on the decision fusion performance in networks with an unknown number of sensors. In performing the decision fusion process, the fusion center applies the Chair-Varshney test; suitably modified to take account of the unknown network size. A closed-form analytical expression is derived for the error probability of the modified fusion rule. It is shown that reducing the censoring probability, i.e., allowing a greater number of sensors to transmit their decisions, does not necessarily improve the decision fusion performance. Rather, there exists a certain censoring probability threshold below which increasing the number of transmitting sensors simply incurs a greater intra-network communication overhead but without improving the global decision performance. Our findings establish that the design of energy-efficient local detection rules should commence with the censoring rate threshold. Hence, it is desirable that the value of this censoring probability threshold be known in advance. Accordingly, the present study proposes an efficient method for identifying the censoring probability threshold value and determining the corresponding local censoring rule.