Harvested Energy Prediction Technique for Solar-Powered Wireless Sensor Networks

Abstract

Solar energy harvesting (EH) is one of the best promising approaches toward perpetual network operation, and it is implemented in various regions of interest (RoIs). However, saving external energy is the essential prime factor in any embedded sensor with finite storage capacity. Generally, the energy conversion rate of the solar system is too fast due to various environmental conditions. Besides, ambient resource energy is noncontrollable, and rechargeable battery only operates in outdoor ecological systems. Frequent environmental fluctuation in their prediction is imperative for initial energy control. Considering the challenges mentioned above, in this article, we propose a modified PROfile energy (Pro-energy) prediction technique to control unnecessary errors in solar-based harvesting systems related to the sensing devices, which estimates the most similar profile-based energy observation in previous time slots. Our proposed method uses prior energy measurements to show future energy status in the respective time slots. Experimental observations on various performance matrices validate that the modified Pro-energy prediction technique exhibits more promising and superior performance than existing EMWA, weather-conditioned moving average (WCMA), and Pro-energy methods.