Automated zone-specific irrigation with wireless sensor/actuator network and adaptable decision support

• Precision irrigation based on the “speaking plant” approach and accurate measurements.
• A modular wireless sensor node used in reconfigurable autonomous sensing networks.
• A proactive closed-loop computing model through an adaptive decision-making layer.
• Machine learning to determine accurately significant thresholds of plant-based parameters.

Precision irrigation based on the “speaking plant” approach can save water and maximize crop yield, but implementing irrigation control can be challenging in system integration and decision making. In this paper we describe the design of an adaptable decision support system and its integration with a wireless sensor/actuator network (WSAN) to implement autonomous closed-loop zone-specific irrigation. Using an ontology for defining the application logic emphasizes system flexibility and adaptability and supports the application of automatic inferential and validation mechanisms. Furthermore, a machine learning process has been applied for inducing new rules by analyzing logged datasets for extracting new knowledge and extending the system ontology in order to cope, for example, with a sensor type failure or to improve the accuracy of a plant state diagnosis. A deployment of the system is presented for zone specific irrigation control in a greenhouse setting. Evaluation of the developed system was performed in terms of derivation of new rules by the machine learning process, WSN performance and mote lifetime. The effectiveness of the developed system was validated by comparing its agronomic performance to traditional agricultural practices.