Automation of On-Farm Irrigation: Horticultural Case Study  Su Ki Ooi ∗ Nicola Cooley ∗∗ Iven Mareels ∗ Greg Dunn ∗∗ Kithsiri Dassanayake ∗∗ Khusro Saleem ∗∗∗ ∗ NICTA, Victoria Research Laboratory, Department of Electrical and Electronic Engineering, Melbourne School of Engineering, The University of Melbourne, Parkville VIC 3010 Australia (e-mail: skoo@ee.unimelb.edu.au, i.mareels@unimelb.edu.au). ∗∗ Faculty of Land and Food Resources, The University of Melbourne, Parkville VIC 3010 (e-mail:{ncooley, gdunn, kbd}@unimelb.edu.au). ∗∗∗ NICTA, Victoria Research Laboratory, The University of Melbourne, Melbourne VIC 3010 Australia (e-mail: khusro.saleem@nicta.com.au). Abstract: Our planet is facing a serious fresh water crisis and improved management of water resources presents signiﬁcant research challenges. About 70% of the world’s fresh water is consumed by agriculture and water application efﬁciencies are typically low. Improving the efﬁciency of water use in agriculture would deliver substantial economic and environmental beneﬁts. To help address the issue of low water application efﬁciency, wireless sensor network technologies can be combined with automation to support the efﬁcient production of food in a increasingly water limited future. This paper reports on results of combining these approaches to irrigate an apple orchard, where existing irrigation scheduling was predominately based on time consuming manual acquisition of soil moisture data. The system developed allowed fully automated on-farm irrigation based on real-time feedback control which increased economic water use efﬁciency by 73% compared with manual irrigation. It is suggested that the application of real-time feedback automation would dramatically improve economic efﬁciency for irrigators with low water efﬁciency. Whereas, for those irrigators already achieving high water efﬁciency, adoption of the technology provides substantial labour and time savings. Keywords: Agriculture, Environments, Feedback control, Automation, On-off control, Wireless sensor networks. 1. INTRODUCTION As our planet is facing a serious fresh water crisis, improved water resources management presents important engineering and research challenges. UNESCO (United Nations Educa- tional, Scientiﬁc and Cultural Organization) indicates that the world is heading for a water shortage under a no-change management scenario and median forecast population growth (Global Water Situation 2010). About 70% of the world’s fresh water is consumed by agriculture. Recently, in Water Report 2 (2006), it was estimated that our population requirements for water in agriculture will grow to 600km 3 by 2025. The world average efﬁciency of irrigation water use is estimated to be around 40% (Water Report 2003); that is about 60% of the water withdrawn for irrigation is wasted. Improving the efﬁciency of the productive use of water in agriculture would deliver substantial economic and environmental beneﬁts. Moreover, a substantial contribution to a more sustainable use of fresh water resources has important socio-political implica- tions. Recent studies have demonstrated the potential of real- time feedback control principles in irrigation water manage-  This work is jointly funded by Victoria State government, The University of Melbourne, NICTA VRL and Goulburn Murray Water under the Science Technology and Innovation initiative infrastructure grants program - ”Regional and Economic Beneﬁts through Smarter Irrigation”. ment to increase the efﬁciency of water application and this is an area which attracts increased attention, see e.g. Schuurmans et al. (1999), de Halleux et al. (2003), Litrico and Fromion (2003), Mareels et al. (2003a), Dulhoste et al. (2004), Mareels et al. (2005b), Mareels et al. (2005a), Mu˜ noz and Dukes and references therein, Cantoni et al. (2007), Mareels et al. (2007), Weyer (2008) and Ooi and Weyer (2008). To approach this resource management problem, a multidis- ciplinary research team was put together, consisting of re- searchers based at The University of Melbourne, National In- formation and Communications Technology Australia Lim- ited (NICTA) and Goulburn Murray Water (GMW). The team wanted to increase economic return of water ($crop/ML water applied) in the orchard. This could be achieved by a more responsive irrigation system, where sensors in the ﬁeld (either directly or indirectly) indicate the volume of water used by the plants. In order to enable this approach new technology was to be developed. A wireless sensor and actuator network was established, to measure the water needs of plants, and deliver the required water when and where it is needed. The main aim of the project is to support continued growth in these major Australian food industries in the face of an increasingly competitive water market (due to population growth) and ex- tended draught conditions (worst on record over more than a decade). The project was supported through the Victoria state