MAP Estimation of Network-Coded Correlated Sources
In this paper, the authors consider a Wireless Sensor Network (WSN) in which sensors measure spatially correlated data and transmit these data to some data processing sink. Random Linear Network Coding (RLNC) is performed at the intermediate nodes of the network. A Maximum A Posteriori (MAP) estimator is considered at the sink to exploit the spatial correlation between data samples and provide a reconstruction of the data, even if not enough network-coded packets have been received, which usually makes network decoding very difficult. Experimental results show that with the proposed MAP estimator, the reconstruction quality increases gracefully with the number of received packets.