Decision fusion with censoring wireless sensors under Rice fading channels

Abstract

This paper addressed the issue of fusing decisions transmitted via Rice fading channels, with censoring transmission strategies. We derived an optimal likelihood ratio rule based on the channel state information. However, acquiring channel information may be too expensive for resource-constraint sensor networks. Under this circumstance, suboptimal likelihood ratio fusion rule is proposed, requiring only the information of channel statistics. In addition, when the signal-to-noise ratio (SNR) of channels is quite low, the reduced forms are also derived. Results from both theoretical analysis and simulations show that suboptimal fusion rules need much less information and computation, which only causes slight performance degradation. Suboptimal fusion rules are practicable for decision fusion systems in resource-constraint censoring wireless sensor networks.