A Performance Evaluation of Solar Energy Prediction Approaches for Energy-Harvesting Wireless Sensor Networks

Abstract

Energy harvesting from the surrounding environment has been a superior way of eliminating the burden of having to replace depleted batteries in wireless sensor networks (WSNs), thereby achieving a perpetual lifetime. However, the ambient energy is highly time-variable and depends on the environmental conditions, which raises the need to design new approaches for predicting future energy availability. This paper presents a performance evaluation and comparison of three recently-proposed solar energy prediction algorithms for WSNs. In order to provide an accurate performance of the algorithms, real-world measurements obtained from a solar panel were considered. Also, the performance characteristics of the algorithms in four seasons –winter, spring, summer and autumn – were demonstrated. To do this, a month in each season was selected for performance comparison, discussing the performance of the algorithms in each season.