3 minute read
Digital twinning in livestock systems and agriculture: Industry 5.0
MAZDAK SALAVATI, READER, SRUC
The use of data and software in livestock and agriculture began in the late 1970s and early 1980s. However, integrating data-driven and digital tools into farm operations has not been practical due to market and cost barriers, until now.
The rise of Internet-of-Things (IoT) devices in the early 2000s has made an interconnected data ecosystem in farming more desirable. One such framework enabling this is digital twin (DT) technology. A digital twin is a detailed digital replica of a real-world entity, allowing a continuous and two-way flow of information between the replica and the original.
For decades, monitoring and managing physical entities like machines or feed batches in real time has been essential in production. Recently, the focus has shifted to creating machine-to-machine (M2M) data ecosystems, where processes can run alongside simulations.
Examples include weather simulations by the Met Office and engine development DTs by Rolls-Royce. However, some argue that without two-way data flow, these are just “digital shadows.” mazdak.salavati@sruc.ac.uk
Regardless of the academic nuance, there is an undeniable need for interoperable data streams in the livestock farming sector that DTs can deliver on.
At SRUC’s Dairy Research and Innovation Centre, we have been focused on a principle that a DT, built on delivering a specific function, has exponentially better chances of not becoming a shadow and taking a leading role in information and asset management.
In recent years, IoT devices have been deployed at scale for the purpose of sensing and monitoring key performance indicators across various scenarios such as farm machinery, sustainability and biodiversity sensors in soil, water and the environment (GHG emissions), animal wearables, camera and computer-vision enabled tasks (e.g. early detection).
Most of these precision farming and livestock farming technologies were developed in isolation without much interoperability capacity at the heart of each technology which has become the main catalyst in developing DT’s for the livestock farming ecosystem.
On the other hand, most farm management software has grown increasingly specialised, especially in the dairy sector, integrated to the benefit of a sustainable business model. Most of these integrations are achieved through in-house development of software that can consume data streams from various sources either directly or via application programme interfaces (API).
The widespread use of open APIs and well documented practices in the industry has hugely affected the ease of integration even between systems that don’t necessarily produce compatible data models like milking parlour software, feed mixer wagon and milk recordings for integration with farm management software.
There is a growing consensus amongst livestock practitioners, data and non-data scientists alike, for establishing system level ontology (description of terms) and data standards for the benefit of the greater integration between different platforms. The data architecture and computer requirements for such applications has gone through similar evolutionary processes while the interoperability challenge has remained the same.
At SRUC, a large body of research has been done under the Digital Dairy Chain banner on the development of functional DTs for both dairy farms and the wider dairy sector stakeholders. The functional approach to DT design has provided a roadmap for a digital continuum spanning dairy farms, affiliated fields and energy consumption profiles as well as operational improvements.
A human-centric design hopes to enable the use of DT technologies for day-to-day management of farm assets as well as serving as a powerful real-time animal breeding and management tools for research purposes.