AI-enabled Vending Machine with Off-grid Energy Harvesting

In remote areas, vending machines face significant challenges in achieving energy sustainability, particularly due to the absence of conventional energy sources and the low frequency of their use. To address this, we propose an innovative AI-driven methodology that leverages cloud map data for accurate solar energy generation prediction, tailored for the unique demands of vending machines in isolated locations. Our approach introduces a two-phase solar energy generation prediction method that merges a sequential-based cloud map model with a solar energy prediction model, further enhanced by the integration of numerical meteorological features for a comprehensive hybrid prediction system. This methodology marks a significant advancement in energy management for vending machines, offering a pioneering application of AI-driven solar energy predictions to ensure operational efficiency in remote areas. When the predicted solar energy is insufficient to meet operational needs, the system can efficiently use the energy of the power grid as a backup to ensure uninterrupted operation while prioritizing the use of renewable resources. By addressing a critical gap in remote vending operations and showcasing the potential of integrating AI with meteorological data, our work contributes significantly to the advancement of renewable energy solutions and operational sustainability.

Collaboration: The project is in collaboration with Aizu CSL, Inc.