Successful product innovations are crucial for trust in robotics across the sector
6 March 2026
Every two months, we introduce a specialist and share an insight into their research and expectations. This time, it’s Jos Ruizendaal, team leader of Greenhouse Automation and Robotics. About gerberas, digital twins, rockwool slabs and aphids: ‘Get into the greenhouse from day one. Don’t wait until everything is fully developed and only then start testing.’
‘This offers so many possibilities.’ Jos Ruizendaal points to a screen showing a realistic 3D view of a plant. That’s the digital twin: a digital copy of a crop. Leaf spacing, leaf size, number of fruits—it is not only visually accurate, but biologically accurate too. It is built with the same software used to create virtual worlds for games, combined with existing calculation models from plant research. The result: a crop that looks digitally the same as it does in the greenhouse.
Virtual harvesting
Ruizendaal shows how the digital twin moves and responds. In the simulation environment, you can even “touch” the virtual plant. ‘That means you can run repeat tests to see how a robot can best harvest a tomato.’ Does it miss a grip once? No problem. You can repeat the virtual harvesting attempt multiple times on the same plant. Just reset the plant, adjust the settings and try again.
In real life, it’s very different. ‘There, you get one chance per plant. If a harvesting test fails, you can’t go back and see whether a small adjustment would have worked. Greenhouse test days are also expensive. So a digital test environment like this is extremely interesting for machine builders.’
The greenhouse in Renkum
Ruizendaal leads the Greenhouse Automation and Robotics team, part of the Vision + Robotics programme at Wageningen University & Research (WUR). In his back garden on the Dorpstraat in Renkum, he has a hobby greenhouse of about eight square metres. Inside, tomatoes, peppers and cucumbers are growing. ‘I do try to get some production out of it. Otherwise there’s no point.’
Sensors measure soil moisture. A small controller manages the water supply. Indoor temperature and humidity are measured too. ‘What I still need to sort out is having the window open and close automatically based on that humidity and temperature.’
Everyone a vegetable patch
He would also love to see everyone in the Netherlands maintain a small vegetable patch: a head of lettuce, spinach, courgette or strawberries. ‘So you can see where your food comes from and what it takes to grow it.’ Nothing you find in the supermarket is guaranteed. ‘We’re so spoiled and so used to everything always being available.’
Line management
Within his department, he manages a fifteen-strong team: a mix of specialists in deep learning, computer vision, robotics, grippers, spectral data and simulations. They work on projects for greenhouse horticulture and open-field crops. He describes his job as a combination of project management, business development and line management. ‘I have to make sure we have enough interesting projects, and that my people can do their work well.’
Gloriosa in Groessen
His fascination with plants is no coincidence. He was born and raised on a nursery. His parents grew gloriosa: a niche crop for cut flowers and houseplants, propagated via tubers—similar to tulip bulbs. The Netherlands had five gloriosa growers, and his father’s business was the largest at two hectares. Stamp-sized, especially compared to an average cucumber grower, who is quickly five times larger.
Outdated greenhouse structures
His father would have loved it if he had taken over the business. But Ruizendaal was more drawn to the technical side than the green side. He also saw up close what it means to be a grower: always busy, seven days a week. ‘You really have to want that life.’ His sisters were not interested either. ‘The greenhouse structures were outdated.’ So the choice was: invest or let it go. His parents chose the latter. ‘They sold the crop. The greenhouse was demolished.’
Agrotechnology
In Wageningen, he chose to study Agrotechnology. ‘Call it mechanical engineering for the agricultural sector.’ After graduating, he stayed with his internship company in Pijnacker: Hortimax, now Ridder. Among other things, they make climate control systems for greenhouses. His focus as an engineer was air circulation, dehumidification and active ventilation for a new type of greenhouse. Both technically beautiful and challenging—but also a daily ordeal commuting from Wageningen. Move to the west of the country? No. ‘I’m more of an eastern-half-of-the-country person.’
Amphibious boats
He moved to Louis Nagel, an importer of agricultural machinery, working as a product specialist. He developed GPS systems for arable farming and was involved with Precision Makers, which developed a fully autonomous tractor without a cab. Later, he joined Conver within the same parent company, providing sales support for mower-collector boats: amphibious boats designed to cut, collect and remove aquatic plants and floating debris from ditches and lakes.
It was not really his world. From greenhouses to fields and waterways—he wanted to return to his own domain. That became possible nine years ago, when he ran into his former WUR internship supervisor, who pointed him to a vacancy at WUR: research into agri-robotics. ‘As if it was meant to be.’
Gerbera harvesting robot
One of his most enjoyable projects in recent years has been the gerbera harvesting robot, developed after a request from the Dutch gerbera cooperative. The reason: a growing labour shortage during harvest. It is labour-intensive work, and it is becoming harder to find the right people. Staff turnover is high and labour costs keep rising. The research started in 2020, during the COVID period. It was financed through a PPS (public-private partnership) arrangement, where part of the funding and risk sits with the government. ‘If this leads to a commercial product, part of that money flows back.’
A stem with a small wound
In the first part of the research, the gerbera itself stood central. Traditionally, the flower is not cut, but plucked by hand. Growers were clear: a robot should do the same. ‘If you cut, you leave a small piece of stem with a wound. Sugary sap flows out, and that becomes food for fungi. Then your plant dies.’
That is why WUR started with a robot that plucks, Ruizendaal explains. And yes, it works—but it is slow. ‘The plucking direction is crucial, and robot eyes from above can’t see that well. If plucking from one direction fails, a person can switch quickly and pluck from the other side. A robot has to “feel” that thinking process and copy it step by step. That takes time.’
Not pluckers, but cutters
So he decided to run cutting tests in parallel. What did they find? The fear of fungi was largely based on old growing conditions. Modern cultivation has changed a lot. ‘In our own trial greenhouse in Bleiswijk and with growers in Mijdrecht and De Lier, we saw no issues with sugars and fungi. So in the second year we changed the grippers from pluckers to cutters.’
Energy crisis
They built a prototype that worked well. The next step was a commercial product. Ruizendaal: ‘At WUR we are good at research and building prototypes, but we are not a machine builder that can launch a commercial harvesting robot.’ Together with growers, he looked for suitable machine builders and held advanced talks with some. Still, the process stalled in 2023. The cause: rising energy prices. Growers’ priorities shifted to reducing energy use.
You need tech-minded growers
This shows how vulnerable innovation is to external developments, no matter how strong the product idea is, Ruizendaal says. Within the cooperative, the dynamics between growers also matter. ‘You really need a group that is tech-minded and actively thinks along. If that forward-looking group becomes too small, you lose the critical mass to continue a project.’ Talks have now restarted with growers and other parties to explore how to give the gerbera harvesting robot a reboost.
Not quite there yet
He knows the gerbera crop is significant, but small worldwide. That makes the business case harder. At the same time, the plant is ideal for robotics: a long, leafless stem with no visual obstacles. ‘But you do need to develop specific components. If you can make modules more widely usable, the robot becomes interesting for more parties.’
Across greenhouse horticulture globally, automation is clearly growing—from harvesting and crop handling to climate control. Yet the feeling has been the same for a few years: we’re not quite there yet. ‘The sector is looking out for a success story. A real breakthrough that everyone involved can build on.’
That “not quite there yet” is mostly about complexity. No plant is the same. The biggest gap is between “we proved it in a prototype” and “we have a commercial product”. A company has to take that step. ‘In business, they may underestimate how much time and effort those last metres take. You keep getting stuck on that’
Learning by doing
What might help. Companies often stay too focused on technology for technology’s sake. Ruizendaal: ‘We’ve seen it a few times: they design a robot on the drawing board, build it in the workshop, and then go into the greenhouse thinking, right, now it has to work.’ Wrong. That’s where development really starts—something he experienced himself during the gerbera robot project.
Go into the greenhouse from day one. Learning by doing. Don’t wait until everything is fully developed and only then start testing. It is practical and sometimes painful. ‘You get feedback from real life straight away—because no plant is the same.’
He points to a potato harvester. ‘If it can harvest a whole field, it will keep driving. But as soon as it has to harvest selectively, it must scan and recognise every single product. That is much more complex.’
Club of 100
Ruizendaal is currently working on six projects. He has already mentioned the gerbera robot and the digital twins. Another large project focuses on phenotyping plant roots on rockwool slabs, commissioned by the Club of 100: a partnership of leading companies and knowledge partners in greenhouse horticulture.
‘We map how roots develop on those slabs,’ he says. ‘The main root, side roots, where they are, and how they are distributed.’ That knowledge can be used in several ways. For breeders, it helps to see how a plant responds to drought and nutrient availability. ‘For suppliers, it’s important to know where roots grow, where wet and dry spots form, and why the distribution is even—or not.’
A standard for rockwool slabs
Often there are three plants on one slab. ‘From a breeding perspective, you can see the differences between those three plants and try to understand where they come from,’ Ruizendaal says. ‘Maybe in the future we should place two or four plants on a slab.’
These measurements provide a lot of information, including the basis for a standard slab setup that everyone within the Club of 100 can use. The research does not stop with rockwool, he emphasises. ‘We also look at other growing media and how roots develop in those.’
Aphids
Another study focuses on detecting aphids in peppers using cameras instead of the familiar yellow sticky traps: half-A4 sheets with glue. ‘Those yellow cards work well, but most aphids don’t fly and don’t end up on traps. To find them, you literally need to be right on the pepper plant—and that’s what we use cameras for.’
That also brings the challenge: to detect aphids, you need to zoom in on the crop. But then you only see part of the plant. The solution is a liquid lens. From a single position, it can take dozens of images, each with a different focus plane. ‘We stitch the sharp parts together into one image where everything is in focus. Then we run detection on that. That gives you a much better overall picture of how many aphids are in a crop.’
Ice Saints
Ruizendaal would love to help the Dutch economy move forward with his team. Successful product innovations are crucial for trust in robotics across the sector. The more complex the interaction, the greater the value. The gerbera harvesting robot is such a product.
His ambition is clear: it should be ready by 2029—assuming external conditions remain favourable. ‘If that works, robotics in harvesting could really accelerate,’ he says. And yes, after the Ice Saints, he will be back in his greenhouse behind the house regularly, until the end of October, to see how his vegetables and other crops grow. Maybe he will come up with a few new ideas.
PPS and Europe
Jos Ruizendaal’s team works on projects for industry, but also on European and PPS projects: public-private partnerships between government and companies. Sometimes the question comes from the sector; sometimes an idea starts within Wageningen University & Research itself. In that case, they look for companies and knowledge partners to build a consortium and secure the right funding.
He does not see specific research trends in topics or approach. What does change is the pace. Companies increasingly ask for faster results. Within Vision + Robotics, powered by WUR, both short-term and multi-year programmes are possible. Depending on the question, it can sometimes be possible to use a technical solution within a few months. At the same time, anyone who truly wants to develop new knowledge quickly ends up in more complex (and therefore longer) research.
