In collaboration with Balanced Engineering LLC, Fraunhofer IESE developed a concept to improve and advance the testing and development of autonomous agricultural vehicles.
Last modified:
What it is about:
Increasing the efficiency of agricultural vehicles with the help of FERAL
The challange:
Replacing expensive field tests for autonomous agricultural machinery with simulations
The support:
Virtual validation with FERAL increases the performance of tractors
The result:
New test environment drives efficiency and sustainability in agriculture
Your benefits:
Fraunhofer IESE as your partner for validation and simulation
Together with Balanced Engineering, a consulting company for agricultural machinery manufacturers, the Fraunhofer IESE team implemented the project “Virtual Validation and Development for Autonomous Agricultural Vehicles” in order to optimize the development and testing of autonomous agricultural vehicles. The project included the creation of a virtual simulation environment with the “Car Learning to Act” simulator (CARLA), which simulates realistic agricultural scenarios. By integrating different sensors, ECU and actuator models, autonomous tractors could be tested under different weather conditions and in different obstacle situations. The simulation framework FERAL, developed by the experts at Fraunhofer IESE, enables the connection of multiple simulators and models to validate the decision-making and performance of the vehicles. Different scenarios – from optimal visibility conditions to extreme weather conditions – were simulated in the project with Balanced Engineering to evaluate and improve the robustness and efficiency of the autonomous systems.
Balanced Engineering approached Fraunhofer IESE with the challenge of creating an efficient and realistic test environment for autonomous agricultural vehicles. The goal was to validate and optimize the performance and safety of these vehicles under different conditions – without having to rely on costly and time-consuming field tests.
Fraunhofer IESE supported Balanced Engineering by developing a comprehensive virtual simulation environment based on the CARLA simulator. This environment made it possible to simulate realistic agricultural scenarios and integrate various sensor, ECU, and actuator models. The FERAL tool played a crucial role in this by connecting different simulators and models and allowing flexible, event- and time-controlled simulation.
This enabled the project team to implement automated test pipelines and ensure the continuous integration and validation of the autonomous systems. By simulating different scenarios – from ideal conditions to extreme weather conditions – the performance of the tractors in terms of object recognition, decision-making, and safety mechanisms could be comprehensively tested and improved. The detailed evaluation of the simulation results, including analysis of processor load and sensor performance, helped to optimize the vehicle architecture and ensure efficient data processing and transmission.
The project resulted in a robust virtual test environment that enables autonomous agricultural vehicles to be efficiently tested and optimized under a variety of conditions. The methods and tools developed made a significant contribution to improving the safety and performance of autonomous tractors. In the future, Balanced Engineering plans to further utilize the knowledge and technologies gained to advance the development of autonomous agricultural systems – and thereby increase efficiency and sustainability in agriculture. The collaboration with Fraunhofer IESE laid the foundation for further innovative projects in the field of autonomous vehicle technology.
This is what Rick Weires, Partner at Balanced Engineering LLC, says:
Fraunhofer Institute for Experimental Software Engineering IESE