Task Scheduling Based on Weather Forecast in Energy Harvesting Sensor Systems

Energy harvesting capacity facilitates the perpetual operation of wireless sensor networks. In addition to battery-based systems, harvested energy significantly varies depending on the deployed environment. Therefore, the system must schedule its workload based on current energy reserves and predictions of additional energy. In this case, the efficiency of scheduling is highly sensitive to the energy harvesting prediction. This paper explores task scheduling using prediction based on weather forecasts in energy harvesting systems. Energy prediction based on weather forecasts gives accurate predictions for medium timescales (hours to several days). This paper proposes and demonstrates a scheduling strategy based on weather forecasts. Experiments reveal a 35%-45% improvement.