Nowadays, the word ‘autonomous’ (or a synonym thereof) is being used in conjunction with greenhouses everywhere you look – in magazines, on websites and at trade fairs. Everyone’s telling you that your greenhouse must become data-driven! In fact, if you’re just a ‘regular’ grower who walks around the greenhouse doing what feels right, you really are lagging behind. If you believe the hype, you should be using data, algorithms and AI – and the relevant greenhouse technology is already available.
But if you ask me, there’s an element of risk involved. I’m a big fan of further automation and striving towards the autonomous greenhouse. After all, it’s a little absurd to continue to rely on green fingers and instinct when there’s so much money at stake. But although I’m in favor of autonomous greenhouses, what concerns me is that it’s increasingly becoming a marketing-driven battle rather than a technology-driven one.
There’s a steady stream of articles and information promising all kinds of interesting possibilities. There’s also a steady stream of research proposals or findings relating to the potential of all this new technology. But when you dig a little deeper, you often discover that things are actually less clear-cut. “The autonomous greenhouse is entirely managed from a control room!” So how can we spot diseases in such a greenhouse? And how can we evaluate labor, for example? That’s when things often turn out to be less autonomous after all, with the trained eye of the grower still being essential. I believe we’re actually making good progress but the terminology used leaves too much room for interpretation.
The automotive industry had a similar problem, which resulted in the introduction of a 6-level standard for classifying automated vehicles: the SAE. This standard includes simple yet clear descriptions of what you can expect of a car at each level. These levels are also increasingly being referred to in car magazines, for example; a review might state that the car has level-2 autonomy on the motorway. In other words, everyone is clear on what the car can do autonomously and also what drivers are still required to do themselves.
Now that the autonomous greenhouse is such a hot topic, wouldn’t it be great to develop a similar classification system as quickly as possible? That would enable you to evaluate and compare levels of autonomy, in order to know precisely what you could expect from an automated greenhouse and also when you would be required to do some good old-fashioned growing.