Ray Tracing-based Light Energy Prediction for Indoor Batteryless Sensors

Abstract

Light energy harvesting is a valuable technique for batteryless sensors located indoors. A key challenge is finding the right locations to deploy sensors to provide sufficient harvesting capability. A trial-and-error approach or energy prediction method is used as the solution, but existing schemes are either time-consuming or employing a naïve prediction mechanism primarily developed for outdoor environments. In this paper, we propose a light energy prediction technique, called Solacle, which accounts for various factors in indoor light harvesting to provide accuracy at any given location. Exploiting the ray tracing technique, Solacle estimates the illuminance and the luminous efficacy of light sources to predict the harvesting capability, by considering the spatiotemporal characteristics of the surrounding environment. To this end, we defined the optical properties of a space, and devised an optimization approach, specifically a gradient-free-based scheme, to acquire adequate values for the combination of optical properties. We implemented the system and evaluated its efficacy in controlled and real environments. The experiment results show that the proposed approach delivers a significant improvement over previous work in light energy prediction of indoor space.