Abstract
Lighting has been identified as a priority energy need for people who lack access to clean, safe, sustainable and affordable electricity access. This study presents a smart, solar-powered LED system integrated with a Dynamic Energy Management System (DEMS) to optimize energy allocation for lighting and study time. Designed for rural South Africa, where about 16.6 million households experience energy poverty, the system leverages machine learning (ML) to predict battery charge time based on weather conditions and academic performance based on study hours. A portable smart LED cube was introduced to ten high school students in Xigalo village, significantly increasing their study time (optimized at 9.46 hours) and improving their academic performance from an average of 52.2% to 66.6%. By harnessing solar energy for lighting and cognitive benefits, this AI-driven solution demonstrates its potential to bridge educational inequalities caused by energy poverty.