Reconstruction of Correlated Sources with Energy Harvesting Constraints

In this paper, we investigate how to minimize the distortion in the reconstruction of correlated sources. We consider a communication scenario where a sensor node is capable of harvesting energy from the environment and where the Fusion Center (FC), in order to exploit correlation, uses past observations as side information for decoding. We provide a convex formulation of the problem and derive the optimal transmission policies (i.e., power and rate allocation). We also propose an iterative procedure based on the subgradient method by means of which a solution can be iteratively found. Interestingly, each iteration entails the interaction (coupling) of a directional waterfilling and a reverse water-filling schemes. Numerical results are provided in order to illustrate the impact of correlation in the resulting transmission policies.