Students design robots for flower growers

Around 150 master’s students in Robotics at TUDelft presented their end of year project this week: a robot designed to support flower growers in their daily work. The robots help monitor flowers, detect pests and measure greenhouse conditions, and can also communicate with greenhouse staff. The project marks the end of their first year and was assessed by a jury.

The robotics hall of the Faculty of Mechanical Engineering has rarely looked so colourful. Artificial dahlias are arranged throughout the space, giving it an almost serene atmosphere. Yet that calm is deceptive: between the flower beds, students move around with tense expressions, making last minute adjustments and hoping their robot will perform exactly as intended.

Efficient flower cultivation
“The theme this year is floriculture,” says assistant professor Chris Pek. “It is a global industry in which the Netherlands plays a leading role. About half of the world market is located here, with a production of roughly 200,000 flowers per hour.” That success is partly driven by innovative technology, which helps growers continuously optimise their processes.

Identifying pests and growth issues
Within this context, students, acting as employees of the fictional MDP Inc. Consultancy Firm, were given eight weeks to develop a robot for use in a greenhouse. These robots are intended to contribute to more sustainable cultivation, improve efficiency and reduce labour intensive tasks.

Data collection lies at the heart of the assignment. The robots measure both the climate and the condition of the flowers. Via QR codes placed on the tables, they receive information on factors such as temperature, humidity and CO₂ levels. A range of challenges has also been built in: simulated pests, variations in flower height, unusual colours and even empty spots to mimic growth problems. It forces both the robots and their makers to stay sharp.

Using this data, the teams create a digital replica of the greenhouse, known as a digital twin. Based on this model, the robot can provide advice to staff, for example on overheated zones, overly dry flowers or the overall condition of the crop.

Systems Engineering: from greenhouse needs to robot validation tests
To better understand their design context, many teams visited a real flower grower early in the project. “That experience helps them develop robots that truly match the needs of the people working in the greenhouse,” says Pek. “For instance, by focusing on intuitive human robot interaction, such as robots that communicate through LED signals or point out flowers using a robotic arm.”

This year, students are working with a new platform for the first time: the MIRTE Master Version 2.0. They also use modern AI models in their designs. While this expands what is possible, it also makes it more challenging to ensure the robot behaves exactly as intended. That is why there was additional emphasis on systems engineering, teaching students to view their robot as an integrated system and to verify that it meets all requirements.

“Systems Engineering provides methods to understand and analyse complex design problems, enabling us to focus on what matters in the end: delivering positive value to the stakeholders, from the greenhouse operator to our society at large”, says Carlos Hernandez Corbato, associate professor lecturing in the course and Chair of TU Delft Systems Engineering Platform. Thanks to the support of our colleagues in the Systems Engineering Platform, we are introducing more of these methods into the MSc Robotics and other programs at ME and TU Delft.

Source: TU Delft

Trending Products