Building a Sky Shepherd for the future of agriculture

Abstract

Sky Shepherding is a concept whereby moving tasks are performed by an autonomous airborne agent capable of responding to animal behaviour and improving animal welfare outcomes. To achieve this outcome, a series of experiments investigated the potential welfare impacts of using a purpose-built Sky Shepherd compared to a trained sheepdog. Data was collected using a bespoke system of mobile phones, heart-rate sensors, and edge computing devices. In this article, we present the findings of the experiments, specifically that a Sky Shepherd has the potential to foster improved animal welfare on farms but does not significantly reduce the physiological stress of Merino sheep during mustering. We outline the system requirements developed to build a Sky Shepherd and supporting sensor network and the future directions to support improved performance. Finally, we present the potential challenges to using high data-rate transfer systems in remote areas and possible future mitigations to support autonomous muster capabilities.