Abstract
Distributed estimation in energy-harvesting wireless sensor networks is examined in this work. Here, each sensor takes a local measurement of the common parameter of interest and forwards a scaled version of it to the fusion center through orthogonal channels. The energy available for transmission at each sensor is converted from ambient energy, whose arrival is random. Two analog forwarding transmission schemes, clipping avoidance and best effort are proposed. Based on the information of whether the sensors transmit, we propose three transmission-reception schemes, transmission unaware clipping avoidance (TUCA), transmission aware clipping avoidance (TACA) and transmission aware best effort (TABE) schemes. In TUCA and TACA, each sensor transmits only when its required transmission energy is less than its available battery energy. Besides, in TABE, each sensor transmits regardless of its available battery energy, in which case, clipping errors may occur. The information of whether the sensors transmit is known by the fusion center in TACA and TABE, but not in TUCA. The maximum-likelihood estimator is adopted at the fusion center and is derived based on the statistics of the energy arrival process. The transmission policy parameters of TACA are sub-optimized by average of mean square error bound. The effectiveness of our proposed schemes is demonstrated through Monte Carlo simulations.