Customizable Platform for Autonomous Agriculture Research

点击翻译按钮获取中文摘要

One of the greatest challenges we face globally today is food scarcity. Amplified by the impact of climate change, access to cultivatable land is increasingly difficult, and there is an obvious and immediate need for innovation in the space of effective agricultural practices. One of the prevailing strategies in this space is the area of small-scale vertical farming. With the advent of powerful autonomous system technology, new methods for automated agriculture - on the small-scale vertical farm setting - is ripe for innovation and real-world impact. This proposal aims to bring two powerful research areas - evolutionary computing and automated planning - to the setting of autonomous agriculture. By viewing the range of possible sensors and actuators available for indoor vertical farms as an autonomous system's environment and capabilities (respectively), we will be able to explore how modern artificial intelligence techniques can be leveraged to optimize crop yield in a controlled environment. An overarching aim of the proposal is to enable the full exploration of this setting by having as many aspects of the environment sensed and controlled as possible with modern indoor farming technology. We are proposing a single unified platform that has sixteen individually controllable growing zones for indoor aquaponics farming. All will have a shared water and nutrient system, but the isolated growing environments will allow for controlled experimentation in automation. The platform has been designed in collaboration and consultation with the Queen's Vertical Farm Team, which has established expertise in constructing indoor vertical farm solutions. The proposed expenditures are for the components necessary for the platform and would represent a one-of-a-kind research testbed for autonomous agriculture. The platform will be used to highlight the research of two research programs at Queen's University in the School of Computing. Roughly 45 HQP yearly will have direct access to conduct research with the vertical farm, and countless others globally will have the opportunity to explore the datasets produced as a result of the regular operation of the platform - releasing the data under a permissive academic licence is an objective of the project. The platform will allow the existing diverse HQP teams to exhibit their research and provide students with a very competitive skillset of working with autonomous systems; all the more crucial given the growing demand for autonomous agriculture expertise to address the global need for solutions in this space.