As part of the CSIRO Next Generation Graduates Programs, the AgriTwins Next Generation programme represents a transformative approach to addressing the myriad challenges faced by the agricultural sector today. By seamlessly integrating advanced cybersecurity measures with emergent data-centric twin technologies, we aim to tackle sustainability concerns, optimise efficiency, and ensure resilience in our agricultural practices.
This program leverages cutting-edge digital twin and quantum twin methodologies, offering real-time operational data insights, precision farming capabilities, advanced greenhouse emission management, and proactive agronomy solutions. Our vision is not merely to adapt to the digital era but to proactively shape the future of agriculture, making it resilient against both current and unforeseen challenges.
The AICF Centre’s AgriTwins program will be offering up to 15 PhD/Masters of Philosophy scholarships for outstanding research scholars commencing in 2025/2026. Projects can be expanded below to detail the specific information on each digital twin, scholarship, ideal candidate criteria and eligibility requirements.
SunRice have developed an instrument called the PaddyVision®, which is the first of its kind in the world. The PaddyVision® non-destructively measures paddy rice quality at the point where it is received from a farmer and is a global advance in the real-time analysis of the potential milling quality of rice. This project is an excellent opportunity to be at the forefront of deep learning and machine learning while developing PaddyVision® as a global standard in rice quality measurement.
This project utilises publicly available information such as water allocations, dam inflows, historical rainfall, and crop competition, along with internal data including production volumes, yield, paddy prices, quality, consumer demand, sales, and milling capacity. Using this comprehensive data, the project develops advanced algorithms capable of generating predictive scenarios to help determine optimal commercial offerings, thereby maximising returns for growers and shareholders. Additionally, it creates sophisticated algorithms that segment and profile growers based on historical performance to forecast future outcomes. To enhance stakeholder engagement and decision-making, this project is also developing a digital twin system that provides easy visualisation and utilisation of these insights.
Collaboration: This project is being delivered in collaboration with University of Sydney.
This project explores advanced 3D simulation technology to enhance the management and productivity of the vineyard. The project aims to simulate various management scenarios and evaluate their impacts, by constructing a comprehensive virtual model that encompasses the vineyard’s topography, soil properties, plant variety, and environmental conditions. This model will enable functions to predict the outcomes of different irrigation schedules, nutrient applications, and pest control measures on vine growth, fruit yield, and overall vineyard health.
This project will develop a digital twin platform that leverages real operational data captured from a hydrogen production system to optimise agricultural hydrogen utilisation across multiple sectors. By integrating hydrogen production data, energy consumption patterns, and agricultural operational requirements, this platform will create detailed virtual models that simulate optimal hydrogen deployment strategies for agricultural applications such as livestock operations, vineyards, feedlots, but not limited to these. The digital twin will model various hydrogen utilisation scenarios, production scheduling, and distribution strategies to maximise efficiency and return on investment while ensuring cyber-secure data management and system resilience.
This project aims to apply digital twin technology to map and monitor land use in the dairy production region, focusing on greenhouse gas (GHG) monitoring and enhancing environmental sustainability. Utilising data from satellite imagery, drones, sensors, and weather stations, the digital/quantum twin will integrate land topography parameters, soil health data, and microclimate data to provide a dynamic and comprehensive representation of the region. This approach will enable better decision-making and environmental stewardship in dairy production, GHG monitoring, grazing management, enhancing both productivity and sustainability.
Collaboration: This project is being delivered in collaboration with University of Queensland.
Microalgae and cyanobacteria are promising bioresources for food and plant bio-stimulant products. Algenie developed a novel "helical" photobioreactor for growing microalgae and is hyper-focused on getting production costs as low as possible. A robust feedback control system will be implemented to rapidly identify the ideal growing environments for any species being tested and determine the optimal conditions for growth. This project will utilise a Digital Twin platform, which simulates the entire cultivation process in a virtual environment, aiming to expedite the overall workflow significantly, allowing us to iterate through experimental designs and optimise outputs rapidly.
Details of this project are TBC
Use a large volume of historical data on animal breeding and eating quality to test if producers can reduce cattle days on feed (e.g., from 120 to 100) while maintaining eating quality. This project will adopt a data-driven approach to develop models that assess the ideal animal weight based on feed availability.
The preferred candidate should have sound data science background with machine learning experience, including data preparation and cleansing, predictive machine learning model development and training.
This project is to build models on environmental and geospatial data to determine a decision support framework for variable rate application of pre-emergent herbicides. Using a digital twin virtual farming system aggregates field topography, soil parameters, previous operations and satellite imagery to create zones and scenarios that assess best economical and productive rate of herbicides, reducing crop injury and potentially increasing weed control efficacy. This approach determines the bio-economically optimum herbicide rate for specific zones within paddocks while adhering to registered label rates.
Details of this project are TBC
The Cool Soil Initiative (CSI) has already collected paddock-level data across 190 farms over a period of up to 5 years, including farm practices such as pulse crops, pasture rotation, fertilisation application, soil health parameters, and economic data. This project will integrate these datasets into a digital twin prototype model using the historic data from the current Global Smart Farm Network. This model will be enabled to undertake relational analysis of datasets to understand current climate-smart agricultural innovations and the drivers of resilience in the major Australian cropping region to:
The Charles Sturt University Global Digital Farm is a 2,500-hectare farming enterprise consisting of two properties in Wagga Wagga and Orange, supporting a diverse range of agricultural production, teaching, and research activities.
This project will investigate the spatial and temporal relationships among various soil, plant, and weather data, as well as crop management practices within the dryland cropping program. A range of existing crop and plant growth models (APSIM, Yield Prophet, etc.) that utilise these data and relationships will then be integrated to create a spatially enabled, digital twin-based decision support system focused initially on fertiliser management of cereals. With other management issues, crop types, and pastures also potentially able to utilise a similarly structured capability, an emphasis on developing a template-based framework will be employed to streamline future applications.
This project aims to use electronic ear-tag data to forecast cattle availability for sale across different regions. By analysing individual animal age, movement history between properties, and past turn-off patterns, a data-driven prediction model will give early insight into future cattle supply. Forecast potential kill amounts by region, species, and breed using data on young stock. This can help producers with planning and decision-making.
Forecast potential kill amounts by region, species, and breed using data on young stock. This can help in planning and decision-making for producers.
The preferred candidate should have strong data analytics skills (including SQL), Python programming, and machine learning expertise.
This project will curate and develop a comprehensive verbatim dataset to identify operational and product improvement opportunities. It will also design and implement an AI-enabled system that utilises this enhanced dataset to automatically spot areas for operational and product enhancements. Additionally, the project will conduct sentiment analysis, customer feedback tagging and categorisation, root cause analysis, and trend analysis to derive actionable insights.
The preferred candidate should have proficiency in Python, SQL, experimentation with Large Language Models (LLMs), and business analysis—translating the understanding of the customer feedback process into data-driven insights and solutions.
This project aims utilise quantitative biometric and animal data into practical tools to assist livestock producers in detecting, monitoring, and rapidly intervening rapidly to improve and enhance animal productivity and welfare in feedlots and yards, utilising vision AI and data science techniques. The project will utilise animal data to develop vision transfer-based AI models to identify key features that have the most significant influence on animal condition and welfare metrics, such as BRD, pink eye, lameness, aggressive interactions, injuries, stress, movement, weight change, and shy feeding, among others. Subsequently, it will develop vision AI models of feedlot livestock that can simulate the impacts of livestock management interventions on animal welfare and production outcomes to inform the cost-benefit analysis of intervention strategies.