Irrigation Control Using Predicted Soil Moisture

The system performs preliminary action by calculating a predicted soil moisture value that anticipates future moisture levels before the scheduled irrigation time. This allows the system to activate irrigation in advance when the prediction indicates moisture will fall below the threshold, rather than waiting for the threshold to be actually reached. The predicted value is calculated by adjusting the current moisture reading with a historical adjustment factor derived from negative net changes in moisture values, enabling proactive irrigation scheduling that prevents plant stress while maintaining operational simplicity.

The system uses self-service by automatically calculating the historical adjustment factor from its own stored moisture data without requiring external input. The controller retrieves past moisture readings, computes negative net changes over time intervals, and generates the adjustment factor autonomously. This self-calibrating mechanism enables the system to adapt to site-specific moisture dynamics and weather patterns, activating irrigation only when genuinely needed based on learned historical patterns rather than fixed schedules or excessive safety margins.

The system applies partial action by selectively using only the negative net changes from historical data to calculate the adjustment factor, ignoring positive changes where moisture increased. This focused approach captures the critical drying trends without processing all historical data points equally. The controller identifies periods where moisture decreased, averages or determines the median of these negative changes, and applies this partial historical information to the current reading, achieving good prediction accuracy without the computational burden of analyzing complete historical datasets.