Lightweight energy prediction framework for solar-powered wireless sensor networks

This research studies lightweight energy prediction filters for solar-powered wireless sensor networks. A generalized prediction filter is developed from the empirical analysis of several solar intensity datasets. The Array of Beta Coefficients (ABC) energy prediction filter is proposed. A comparison metric is also proposed to evaluate different filters based on their accuracy, storage requirements, and calculation complexity. Simulation results show that the ABC filter has up to 8-fold accuracy improvement over other published filters.