Energy-Efficient Distributed Detection Via Multihop Transmission in Sensor Networks

We investigate three multihop fusion schemes for distributed detection in geographically dispersed sensor networks, multihop forwarding (MF) and Log-likelihood ratio Fusion (LF), which are different in the transmitted messages, fusion rules, and communication structure. Simulation results show that transmission energy is significantly reduced by multihop fusion schemes as compared to direct transmission, with LF outperforming the others. Moreover, it is shown that LF exhibits the most favorable energy scaling law with the network size among these schemes.