From Lab to Field: Agricultural Robotics Innovations Presented at GOFAR Field Day Spain

The DTDA2 project, funded by the Spanish Ministry of Agriculture and the European Union, brings together four universities: the Politécnica de Madrid, the Universidad de Córdoba, the Politécnica de Valencia, and the Universidad de Zaragoza, alongside agri-food cooperatives as representatives of the productive sector. Over eighteen months, the project has organised twelve demonstration events across Spain covering variable rate application, seed fertilisation, spectroscopy in food industry settings, digital irrigation management, advanced precision machinery, and precision livestock. The Field Day robotics session was one of the project's flagship events, co-organised with GOFAR and John Deere.

An Autonomous Sprayer Built for Continuous Field Operations

Adrián Navacerrada presented the Aries 300N on behalf of AGC Technologies, an autonomous spraying robot that operates without an operator using satellite navigation and obstacle avoidance. The machine is built around a hybrid petrol-electric architecture: a small six-litre engine functions not as a conventional motor but as a generator, charging the battery that powers all operations. "This motor does not work like a car engine," Adrián explained. "It works more like a generator. All the torque and force the robot has comes through the electronics." This gives the robot an autonomy of between two and five hours per tank, with maximum fuel consumption of three litres per hour.

The spraying system features a 300-litre tank with eight variable nozzles configurable either in a ring or as lateral bars depending on crop type. Working width is programmable between three and ten metres, the pump operates at up to 60 bars, and flow rate reaches 16 litres per minute, all adjustable remotely. At standard settings the machine covers four hectares per hour, with a theoretical daily capacity of 100 hectares thanks to its ability to work continuously without operator fatigue.

Adrián pointed to four core advantages: greater precision in product application, reduced labour costs, lower environmental impact, and operation without specialised staff. "People no longer want to do work that requires physical effort," he noted. "I think this is the way out we have, to give farmers solutions so they can keep working the way they always have and keep the fields going."

Two Paths to Autonomy: Livestock Feeding and Intelligent Implement Systems

Abel Pozo, Product Manager at Kuhn Ibérica, presented two autonomous vehicle projects that reflect very different philosophies of agricultural automation.

The first, project Aura, is a fully autonomous mixer and feed distributor for housed livestock. Drawing on Kuhn's nearly 200 years of manufacturing experience and more than 50 years in animal feeding, Aura performs six simultaneous functions without human intervention: automated loading of silage, forage or concentrate; ration mixing; weighing and dosing per animal group; distribution; pushing feed back within reach of animals; and full data collection for herd management. "We really saw the need to automate the animal feeding process," Abel said, noting that the system is designed to be "totally adaptable to any moderately developed agricultural operation and fully scalable." Commercially available in France since 2020 with more than 30 units sold, the system is now being extended to other markets.

⇒ More about AURA

The second project, Karl, inverts the conventional logic of autonomous tractors. Rather than making the tractor autonomous while the implement remains passive, Karl gives the implement its own intelligence. "What is particular about this system is not so much the tractor, which is autonomous, but that the implement it carries has a certain level of intelligence that commands the tractor," Abel explained. The implement dictates speed, working depth, and triggers stops in the event of mechanical incidents. "Karl is not a commercialisable product at the moment," he acknowledged. "It is more of a working laboratory for our implements for the future." The system is already designed to operate in swarm configurations of two or three vehicles, programmable from an office or a mobile phone.

⇒ More about KARL

A Spanish Electric Autonomous Tractor Designed for Real Farm Economics

Voltrac is a Valencia-based company with two years of development and a nine-person team drawn from five nationalities. The vehicle on show was the second prototype, used for ongoing development and customer demonstrations ahead of the third and commercial version. It runs without diesel and without a driver, measuring just under five metres in length, 1.4 metres in height and 1.7 metres in width, with 150 horsepower equivalent, a 40 km/h top speed, a rear three-point hitch capacity of 3,000 kilograms, and four-wheel drive throughout.

The economic logic was stated plainly "If we come to market with machines that cost ten times what a farmer already uses, it makes no sense for the farmer and no sense for us either, because nobody will buy it.” Electrification has allowed Voltrac to reduce its bill of materials sufficiently to target price parity with a conventional diesel tractor in the same category, while also simplifying maintenance and reducing mechanical failure risk. The longer-term roadmap includes embedded AI capable of operating without connectivity, generating live vigour maps, and automating implement control. "If it loses connection, it would not matter," he said, "because our embedded artificial intelligence models could control it and optimise the implements it carries."

⇒ More about Thor

A Harvest Assistance Robot That Keeps the Human at the Centre

Ángela Ribeiro of the Centro de Automática y Robótica, a joint centre of the Universidad Politécnica de Madrid and the Spanish National Research Council, presented a fundamentally different kind of robotics proposition: not a machine that replaces the human worker, but one that works alongside them.

The system is a harvest assistance robot developed for grape picking and reconfigurable for other manual harvest crops. Its logic starts from an honest assessment of current robotic capability: human dexterity in harvesting remains beyond what robotic arms can replicate at commercial scale. "Our idea is to deploy a solution where what we have is a transport robot that follows the user," Ángela explained, "so that the user provides the dexterity in the picking and the robot provides the effort and the force." The platform follows the harvester autonomously, carries a continuously weighed harvest box, and communicates with a central system via GPS. When the box reaches its maximum weight of around 20 kilograms, it automatically triggers the arrival of a second robot with an empty box. "All of this happens without any action required from the person who is harvesting," she noted.

In field trials during real grape harvests at Terras Gaudas, the system produced a 50% increase in output from experienced harvesters compared to the same workers without robotic support. The continuous weighing also generates a high-resolution production map of the vineyard. On deployment, Ángela was direct: "We have to go towards a model similar to combine harvesters. People do not normally buy combine harvesters. We need to shift the business towards service delivery." She closed by mentioning an ongoing EU-funded project through which her centre provides free advisory services to companies seeking to integrate robotic technologies.

Technology Has to Solve a Real Problem, Not Just Demonstrate One

Constantino Valero closed the session by turning the question of viability back to the room. The answers converged around a single point: the technology has to make economic sense for the farmer before it can make economic sense for the developer.

Abel Pozo framed it in terms of accessibility. "We cannot have the acquisition cost or the maintenance cost of this equipment become something that complicates life for the user," he said. "The manufacturer also has to secure the return on investment, because the development of this equipment requires a very high investment in both time and money." Voltrac reinforced the point: "If you do not do these things, everything will end up as shelf projects." Ángela Ribeiro drew a practical distinction between platform robotics, amortised across many tasks like a tractor, and task-specific robotics, which requires either very high utilisation or a service model. "You have to demonstrate first that it is really an advantage to integrate the robot," she said, "and then make it easy to integrate."

Constantino's closing observation brought the session back to its starting point: "The technology has to serve the sector and connect with the farmer, be profitable, but because it solves a problem, not simply because it is technology."

DTDA2 Project Session | Field Day Spain, GOFAR | Presenter: Constantino Valero Ubierna (Universidad Politécnica de Madrid) | Participants: Adrián Navacerrada (AGC Technologies), Abel Pozo (Kuhn Ibérica), Voltrac, Ángela Ribeiro (Centro de Automática y Robótica, CSIC)

Listen to this episode on our podcast [Spanish only]