Industrial - Bachelors
ARBORA is a simple, flexible and powerful computer-vision hardware platform designed from the ground up to help meet the needs of SME farmers as they face the challenges of an uncertain future. By enabling farmers to develop data-driven approaches to decision making, we can ensure a sustainable, profitable and secure future for independent agriculture.
The UN highlights that global food systems face immense pressures from population growth, market volatility, and climate shifts – all of which are felt particularly intensely in Australian agriculture due to the harsh environments and challenging labor markets. With agriculture contributing 2.7% to the Australian GDP, employing more than 250,000 people, and with 99% of farms family-owned, supporting these stakeholders and developing meaningful solutions to their challenges is vital. Small and medium farms are the ones most likely to suffer catastrophic consequences of climate change, yet are critically underrepresented when it comes to technology development and adoption despite the potential solutions that they offer.
ARBORA aims to address the growing gap in technology adoption as a key component to building climate resilience for agriculture. Recognising that the challenges faced by farmers are as unique as they are complex, the project investigated the potential of leveraging data-driven decision making on farms, moving from the traditional trial & error method, which can be costly, ineffective and unreliable, especially in the face of rapidly changing climates.
In order to understand the contexts, challenges and everyday realities of the farming environment, both primary and secondary research methods were implemented. Qualitative research was conducted in the form of semi-structured interviews with both end users (SME farmers) and technology experts. Seven semi-structured interviews were conducted, alongside a site visit in Northern NSW (an area that has been significantly impacted by climate events), and a distributed survey gathering the perspectives of farmers across Australia.
None of it’s making any sense. There’s no real baseline comparative thing this time of year… we just don’t quite understand how [we] could be adaptable and resilient.
Farmer, Northern NSW
[I am] confronted with hundreds of those decisions every single day, yeah? And then how do we wade through that… confusion and all that stuff…?
Farmer, Northern NSW
If we can find ways to build
Farmer & technology Expert, Victoria
diversity of revenue… and
we’re doing good for the actual
environment and for nature, that’s where
everything’s going
The research unveiled that the challenges faced by farmers are incredibly complex and unique – sometimes even changing from meter to meter or field to field. This is a huge inhibitor for technology development due to the challenges of scale and specificity. Therefore, instead of developing direct solutions to specific problems, an opportunity to help facilitate the decision-making and experimental processes of farmers was identified.
Moving from the traditional “trial & error” process, which can be costly, ineffective and unreliable (especially in the face of changing climates), a new model of decision-making was proposed, with data at its core. Building data foundations and accessibility allows a shift to more informed practices, alleviating the complexities of farming-based problem-solving. Gather, interpret, decide and act, all correspond with key stages of the problem-solving process and can be transposed across many farming contexts.
In order to best meet the needs and challenges of the end users, an overall design philosophy was developed and its characteristics carried through to the final outcome.
Due to the intense pressures and conflicting priorities of the day-to-day activities of farmers, the outcome must be understandable, intuitive, and accessible. This can be quantified by minimizing onboarding hurdles and required expertise, minimizing interactions and ongoing effort.
The environments and challenges in the farming context are fluid, complex and uncertain. Flexibility and adaptability of the solution will allow the outcome to continue to provide value to users as their requirements and needs shift. Felxibility also maintains autonomy, increasing the accesibility of the solution.
Providing value to the users is of utmost importance, so the system must provide powerful and useful outcomes. Effective, pragmatic and value-driven innovations are key.
Over a 6 week period, several concepts for were developed using traditional Industrial Design methods, including iterative sketching and concept development, model making, CAD & visualisation, as well as a final high-fidelity aesthetic model.
Designed to be a bridge between data-science and the everyday user, ARBORA is algorithm agnostic, allowing for flexibility and development in the spaces of computer-vision and machine learning. As new or optimised models are released, users can update their system accordingly, even changing models over the course of seasons as their key indicators or points of interest change.
By building on a flexible hardware platform, the system can be upgraded as processing power and data storage capacity improves, mitigating tech-obsolescence, maximising investment value and improving life-cycle sustainability for the hardware components.
The two 16MP high-resolution cameras on the side of the system are used to capture high-detail images of the farming environment. This allows for detailed assessment of plant health through indicators such as leaf colouration, weed or pest identification, and biodiversity counts.
The front camera system is optimised for VSLAM (Visual Simultaneous Localisation and Mapping) processes, in order to support the GPS system for precise localisation under challenging environments. VSLAM relies entirely on the visual data, creating a fail-safe if signal or connectivity is lost. Localisation allows for the data to be captured and mapped according to its real-world location, providing more meaningful insights for farmers to understand performance across their farm.
To further support the localisation, a GPS module, accelerometer, and gyroscopes are all included in the main computation stack. These allow the system to better understand and track its position in space, minimizing the chance of losing localisation in the challenging environments of farming.
ARBORA has integrated the battery system with the data storage, meaning that farmers can simultaneously charge their system while transferring the recorded data with no extra effort.
The E-ink screen is a low-power solution that maximizes visibility in harsh lighting environments, while communicating important information while in the field.
ARBORA is designed to fit within the everyday activities of farmers, attaching to various pieces of equipment and accompanying farmers through their land. Set up and recording is as simple as sliding the system into the mount, turning it on, and pressing record. ARBORA then localises according to GPS or WIFI signals and the VSLAM inputs, and begins to record information as it moves through space.
At the end of the recording session, the farmers can either completely remove ARBORA from its mount, or opt to only remove the battery for charging and data transfer.
Farmers are constantly balancing priorities and making hundreds of decisions every day, so ARBORA is designed to minimize the complexity of the interaction in order to fit seamlessly and effortlessly into their every-day practice.
An E-Ink screen displays the vital information relating to the recording status and history of the device, with a simple 3-button navigation system allowing for initiative interactions.
The display is designed for maximum visibility from a distance, with the recording screen being mostly white, while the menu, error and processing screens are black. This allows the farmers to see at a glance if the system is recording, or if further investigation is required.
A removable battery allows for in-the-field power swaps, minimizing downtime. The system can also be charged via the USB C port when mounted to a powered vehicle such as a tractor or ute.
Farmers can keep track of battery and data systems through the ID number inscribed on each component.
Simplifying the data transfer processes was deemed as crucial for this project, in order to minimize the perceived and ongoing efforts associated with data collection. By integrating the data storage in each battery module, farmers can simultaneously charge the battery while transferring the insights.
Data without interpretation has very limited utility, therefore the ARBORA interface presents insights, not raw data. By simplifying the information to key indicators, with drill-downs into the specific information, farmers can get an effortless understanding of their key indicators without having to go through extensive data-interpretation processes.
ARBORA also focusses on being descriptive, rather than prescriptive, allowing farmers to maintain autonomy on how they manage their farm.
The impact and value that ARBORA provides can have far-reaching effects across the agricultural industry. Below, the impacts and value for farmers and the broader system are explored.
For farmers with extensive experience in the industry, but minimal experience in new technologies, the new-found data-driven approach can allow for deeper and more consistent insights into farm performance and overall health, facilitating more informed and longer-term decision making practices.
For example, data on plant health and performance can provide insights into new species and their performance on the farm, allowing for higher confidence in selecting ongoing crop varietals. Tying yield and performance data with the comparative performance of other varietals, weather and overall seasonal farm health can demystify and derisk the selection of crops for future seasons.
As the agricultural industry undergoes a rapid generational shift, ARBORA supports digital natives with integrating data-collection into their everyday tasks, and facilitates their learning as they adapt to new realities of climate change and market dynamics.
The platform allows these farmers to learn quickly about new farming practices, reducing implementation and optimization time. As farmers adapt to the uncertain future, diversification, regeneration and cover cropping can all be supported through insights provided by the ARBORA system.
For farmers who have already integrated technology and innovation into their practice, ARBORA can build on this foundation and tie in directly with established technologies in the farming space.
GIS, aerial photography, soil testing and mapping can all be integrated with the visual information and mapping provided by ARBORA, which is vital to mitigate the abstractions that occur with established technologies. ARBORA addresses the gap of “ground-truth” farm performance, which is still left to farmers to assess.
Building data literacy and data history for farms in the immediate future will help to understand how our ecological and food production systems are changing, and where support is needed. This can be used on a micro-level inside the farm gate, as well as extrapolated to the community and policy levels, where regions may undergo significant ecosystem changes as climates shift and cropping regions change.
The potential of ARBORA lies far beyond data gathering and interpretation. By selecting a visual-first system, it unlocks the power of technologies such as Gaussian Splatting and Photogrametry. The front 3 cameras can be used to begin building 3D spatial data on each farming environment which can be used, for example, to build digital twins.
Excitingly, it also has implications in autonomous robotics, where the extensive 3D data can be utilised to build versatile training data based on ground truth and actual farming environments (with real-world obstacles, weather, and variability) for navigation systems.
Having the capability and the data history to train an autonomous system on a user’s specific and unique farming environment before deployment, significantly reduces the barrier of entry to an autonomous agricultural future.
The core purpose of this project was to help facilitate climate resilience for SME Farmers. ARBORA addresses this by empowering farmers to make long-term, informed decisions about their farm management, enabling new farming practices (such as cover cropping or diversification) that may take years to implement with traditional processes, and providing transparency and traceability on farm health and performance. This information can be used to build ecological health, which is tied directly to farm longevity and profitability.
Kilian is an impact-focussed Industrial Designer and Mechanical Engineer who believes deeply in the potential of design, engineering, and business practices to drive positive change across industries and communities. This passion has brought him opportunities to work with multi-disciplinary teams across the globe to develop innovative and implementable solutions to our most pressing challenges.