Real-time Monitoring and Forecasting of Ecological Processes (1703.07162v1)

Abstract: The paper introduces a real-time monitoring and forecasting system for ecological phenomena. The process yields a collection of ecological parameters viewed as distributed time series, which are measured by means of wireless network of sensors. The acquired data are preliminary processed and modeled by using complex algorithms in view of prediction. There are three graphical user interfaces implemented within the monitoring and forecasting system: eko-View and eko-Greenhouse (which directly interacts with the process) and eko-Forecast (which estimates the future evolution of some ecological parameters). The monitoring system was effectively integrated in an industrial application dealing with automatic irrigation of a small greenhouse. The forecasting simulation results with real data and a comparative assessment of predictor performances are presented in the end.