A Weather-Condition Prediction Algorithm for Solar-Powered Wireless Sensor Nodes

Abstract

Recently, Wireless Sensor Networks (WSNs) begin to use the solar energy. But the energy from the environment is usually unstable, An efficient solar prediction algorithm should be studied. In this paper, a novel solar energy prediction algorithm, namely Weather-Conditioned Selective Moving Average (WCSMA), is proposed by using the trend similarity of energy harvesting and the classification of sunny and cloudy days. The simulation results show that the relative mean error of WCSMA algorithm only around 10%, and it is much lower than the Exponential Weighted Moving Average (EWMA) algorithm which is widely used now.