Self-Driving Tractor Navigates Groves
Self-Driving Tractor Navigates Groves


An autonomous tractor with three different steering modes can drive in straight lines, make efficient turns, and shift modes in response to what it sees as it navigates through an olive grove.
The adoption of autonomous robots for farm applications is on the rise, and many of these robots rely on electric power and differential steering, which limits what they can do in demanding environments. Determined to design a better agricultural robot, a team of researchers from the University of Córdoba in Spain has developed an autonomous tractor. The diesel- and hydraulically-powered tractor, specifically designed for fruit tree plantations, incorporates a mechanical steering system with dual steerable axles.
“We wanted to tackle a real-world challenge where our mechanical and electronic engineering skills could converge,” said Sergio Bayano Tejero, an industrial engineering researcher with the university. Bayano Tejero and Rubén Solá Guirado, a mechanical engineering researcher at Córdoba, led the project.
“The key moment came when we tried to purchase an autonomous diesel-powered tractor and found that the national market was very limited, with extremely high prices and vague, inaccessible information—something that does not help knowledge development within the university setting,” Bayano Tejero said. “That is when we decided to try and build something close to a real commercial solution ourselves.”
Bayano Tejero and Solá Guirado, with support from the Rural Mechanization and Technology Group at the University of Córdoba, designed the vehicle to have three different steering modes. These modes were: Front (or rear) steering, in which a single axle turns; inverse front-rear steering, in which both axles turn, for a smaller turning radius; and hybrid steering system in which the front axle turns, and the rear axle also turns, but to a lesser extent, enhancing straight-line performance.
These modes allow the tractor to drive in straight lines, make turns efficiently, and shift modes in response to what it sees as it navigates. Two independent self-leveling axles with steerable wheels make control of the tractor much more versatile.
Technologies integrated into the tractor for autonomous navigation include two LiDAR sensors (one in the rear and one in the front), an inertial unit (measures acceleration and inclination), a digital compass to monitor the tractor's direction, and a high-precision GPS system. All programming was carried out using the ROS (robot operating system) environment.
To evaluate the different modes, the tractor was tested in a dense olive grove, where it was determined that the inverted mode was optimal for completing turns precisely and the hybrid mode was most suitable for straight sections. The tractor is able to shift between its different steering modes as needed.
“The results obtained can help establish the necessary control method for autonomous machinery of similar fruit trees,” Bayano Tejero said.
Bayano Tejero and Solá Guirado also showed that innovation does not require expensive, customized, or specialized parts. For example, they reused and adapted off-the-shelf mechanical components to achieve a flexible, customizable platform. From a mechanical standpoint, they developed a compact steering linkage system driven by a linear actuator that was able to maintain precision without needing expensive encoders. The modular design also allows for different tool attachments, which makes the tractor versatile for research and small-scale farming applications.
“But above all, what truly sets our project apart is that we developed a hybrid steering mode that, to our knowledge, does not exist elsewhere,” Solá Guirado said. “In this mode, the vehicle steers both its front and rear wheels in the same direction, but with the rear wheels turning only half the angle. This allows the vehicle to correct the small deviations that autonomous systems can experience when navigating between waypoints. As a result, the vehicle can move in a parallel path while gently adjusting its heading to smoothly return to the planned trajectory.”
“We integrated mechanical design, embedded systems, and control logic into a fully autonomous tractor that we built ourselves from scratch, but it required a lot of testing in field to validate our advances,” Bayano Tejero said. “Coordinating all these systems, especially ensuring reliability in field conditions, was a major hurdle. Communication between sensors, actuators, and control logic, especially under unpredictable terrain conditions, also required extensive testing and fine-tuning.”
The team was also surprised by how much could be achieved using open-source tools and resourceful engineering.
Another unexpected aspect was how carefully the mechanical and electronic components had to be co-designed—seemingly minor design choices in the chassis, for example, had big consequences on electronic integration and sensor accuracy.
“A large number of software modifications were required to adapt navigation from one set of conditions to another,” Solá Guirado added. “This makes it very difficult today to have a truly versatile solution that works for any type of orchard or plantation. As a result, we are seeing more highly specialized solutions in the market. However, we are confident that, with greater funding to hire research personnel, we could generate much broader and more open knowledge—enabling companies to develop their own commercial solutions.”
Designing a robust and modular chassis that supports autonomous operation in agricultural environments was a challenge. The team also had to design custom steering and drivetrain systems compatible with electronic actuation, while ensuring serviceability and cost-efficiency. The integration of a diesel propulsion system with a tandem hydraulic pump setup provides great versatility in operation management, “allowing us to perform a wide range of tasks with independently controlled functions,” Bayano Tejero said. “However, this also requires more space and demands careful optimization of the vehicle’s overall size. The project demanded a lot of interdisciplinary thinking, especially in integrating motion mechanics with sensor data and control loops.”
The engineers are currently working on improving the autonomy level, including more advanced obstacle detection and adaptive path planning. In the longer term, they envision this platform evolving into an open-source educational tool and a base for precision agriculture applications, especially in regions with limited access to high-cost commercial solutions.
Several components of the control system, especially the low-cost sensor fusion algorithms and actuator control strategies, could be adapted to other fields like warehouse robotics, last-mile delivery vehicles, or even mobile inspection platforms in industrial environments.
“What really made this project special was the collaborative spirit and hands-on learning,” Solá Guirado said. “It demonstrated how engineering is not just about iSolá Guiradoted disciplines, but about bringing them together to solve practical problems. Despite our small team and tight budget, we have built something functional, adaptable, and meaningful—and that is something that makes us very proud. This project proves that, with adequate funding, great things can be achieved.”
Mark Crawford is a technology writer in Corrales, N.M.
“We wanted to tackle a real-world challenge where our mechanical and electronic engineering skills could converge,” said Sergio Bayano Tejero, an industrial engineering researcher with the university. Bayano Tejero and Rubén Solá Guirado, a mechanical engineering researcher at Córdoba, led the project.
“The key moment came when we tried to purchase an autonomous diesel-powered tractor and found that the national market was very limited, with extremely high prices and vague, inaccessible information—something that does not help knowledge development within the university setting,” Bayano Tejero said. “That is when we decided to try and build something close to a real commercial solution ourselves.”
Bayano Tejero and Solá Guirado, with support from the Rural Mechanization and Technology Group at the University of Córdoba, designed the vehicle to have three different steering modes. These modes were: Front (or rear) steering, in which a single axle turns; inverse front-rear steering, in which both axles turn, for a smaller turning radius; and hybrid steering system in which the front axle turns, and the rear axle also turns, but to a lesser extent, enhancing straight-line performance.
These modes allow the tractor to drive in straight lines, make turns efficiently, and shift modes in response to what it sees as it navigates. Two independent self-leveling axles with steerable wheels make control of the tractor much more versatile.
Technologies integrated into the tractor for autonomous navigation include two LiDAR sensors (one in the rear and one in the front), an inertial unit (measures acceleration and inclination), a digital compass to monitor the tractor's direction, and a high-precision GPS system. All programming was carried out using the ROS (robot operating system) environment.
To evaluate the different modes, the tractor was tested in a dense olive grove, where it was determined that the inverted mode was optimal for completing turns precisely and the hybrid mode was most suitable for straight sections. The tractor is able to shift between its different steering modes as needed.
“The results obtained can help establish the necessary control method for autonomous machinery of similar fruit trees,” Bayano Tejero said.
Bayano Tejero and Solá Guirado also showed that innovation does not require expensive, customized, or specialized parts. For example, they reused and adapted off-the-shelf mechanical components to achieve a flexible, customizable platform. From a mechanical standpoint, they developed a compact steering linkage system driven by a linear actuator that was able to maintain precision without needing expensive encoders. The modular design also allows for different tool attachments, which makes the tractor versatile for research and small-scale farming applications.
“But above all, what truly sets our project apart is that we developed a hybrid steering mode that, to our knowledge, does not exist elsewhere,” Solá Guirado said. “In this mode, the vehicle steers both its front and rear wheels in the same direction, but with the rear wheels turning only half the angle. This allows the vehicle to correct the small deviations that autonomous systems can experience when navigating between waypoints. As a result, the vehicle can move in a parallel path while gently adjusting its heading to smoothly return to the planned trajectory.”
Sticking to the Budget
One of the biggest challenges for the team was managing complexity with limited resources. The team was quite small, so each member wore multiple hats—from CAD design and mechanical assembly to circuit design and software development.“We integrated mechanical design, embedded systems, and control logic into a fully autonomous tractor that we built ourselves from scratch, but it required a lot of testing in field to validate our advances,” Bayano Tejero said. “Coordinating all these systems, especially ensuring reliability in field conditions, was a major hurdle. Communication between sensors, actuators, and control logic, especially under unpredictable terrain conditions, also required extensive testing and fine-tuning.”
The team was also surprised by how much could be achieved using open-source tools and resourceful engineering.
Another unexpected aspect was how carefully the mechanical and electronic components had to be co-designed—seemingly minor design choices in the chassis, for example, had big consequences on electronic integration and sensor accuracy.
“A large number of software modifications were required to adapt navigation from one set of conditions to another,” Solá Guirado added. “This makes it very difficult today to have a truly versatile solution that works for any type of orchard or plantation. As a result, we are seeing more highly specialized solutions in the market. However, we are confident that, with greater funding to hire research personnel, we could generate much broader and more open knowledge—enabling companies to develop their own commercial solutions.”
Designing a robust and modular chassis that supports autonomous operation in agricultural environments was a challenge. The team also had to design custom steering and drivetrain systems compatible with electronic actuation, while ensuring serviceability and cost-efficiency. The integration of a diesel propulsion system with a tandem hydraulic pump setup provides great versatility in operation management, “allowing us to perform a wide range of tasks with independently controlled functions,” Bayano Tejero said. “However, this also requires more space and demands careful optimization of the vehicle’s overall size. The project demanded a lot of interdisciplinary thinking, especially in integrating motion mechanics with sensor data and control loops.”
The engineers are currently working on improving the autonomy level, including more advanced obstacle detection and adaptive path planning. In the longer term, they envision this platform evolving into an open-source educational tool and a base for precision agriculture applications, especially in regions with limited access to high-cost commercial solutions.
Several components of the control system, especially the low-cost sensor fusion algorithms and actuator control strategies, could be adapted to other fields like warehouse robotics, last-mile delivery vehicles, or even mobile inspection platforms in industrial environments.
“What really made this project special was the collaborative spirit and hands-on learning,” Solá Guirado said. “It demonstrated how engineering is not just about iSolá Guiradoted disciplines, but about bringing them together to solve practical problems. Despite our small team and tight budget, we have built something functional, adaptable, and meaningful—and that is something that makes us very proud. This project proves that, with adequate funding, great things can be achieved.”
Mark Crawford is a technology writer in Corrales, N.M.

