It’s been a while since I posted an article. Part of the reason is that I’ve been busy with my internship…
Phil, like AMI, is not a person but an acronym – Priva Horticulture Innovation Lab. Back in May I started my internship there, the final part of my master’s degree.
Priva is the market leader in high-tech greenhouse horticulture, with the long-term vision that comes with a family business. I’ve been following them for a few years now. In 2016 I spent two days at Priva and visited UrbanFarmers De Schilde and BrightBox in Venlo.
CEO Meiny Prins’s vision of Sustainable Urban Deltas is compelling too and worth a future article (I’d recommend her recent documentary on this). It’s great and even a bit surreal to be an insider now – we even got to meet Meiny about a month ago and ask her questions.
Last year I spoke with some people from Priva at GreenTech – and a few emails later, I was put in touch with Alastair Monk. Ally was one of the founders of Motorleaf, a Canadian company developing data-driven solutions for greenhouse horticulture and indoor growing. Less than a year ago he helped set up Phil, the autonomous and agile speedboat to work within and alongside the solid-but-slow oil tanker of Priva.
Phil is part of Priva’s plan to become more data-driven and cloud-based in general, especially for horticulture. Priva is sitting one of the largest databases in greenhouse horticulture in the world. Now they want to put it to use to gain a competitive advantage.
Plantonomy
As more and more greenhouses get built, and existing companies grow, the knowledge to run them is in short supply. Managing a greenhouse and controlling its climate are complex jobs. On a typical climate computer, including Priva’s Connext system, there are thousands of settings. Plantonomy simplifies this down to just two main settings, with four additional settings, and makes yield more predictable. The ideas behind it have been in development for over a decade through the knowledge of Peter Kamp, who you could call a plant-whisperer.
Although optimal control is fascinating, there’s a lot to be gained from just simplifying the grower’s job and making things consistent and predictable. Unlike a lot of the trials you hear about through the Autonomous Greenhouse Challenge, Plantonomy is actually running on full-scale greenhouses. Because of its simplicity, one grower can control a far larger area, and the learning curve for new growers is less steep. François, a grower in Mexico, explains below:
Last week at GreenTech Don Kester, Global Account Manager at Priva, was interviewed about Plantonomy as well. Plantonomy can make being a grower a more attractive profession. Not only does it improve and simplify control – it means as a grower you can more easily go on holiday, spend the weekend with your family, and do all these other things young professionals expect to be able to do.
Artificial intelligence in horticulture
‘Artificial intelligence’ is a bit of a vague term. As the saying goes, if it’s machine learning, it’s written in Python – if it’s artificial intelligence, it’s written in PowerPoint.
But AI is everywhere. It doesn’t even have to be anything massively complex like IBM’s Watson or AlphaGo. A Google search, satellite navigation, or recognising a song with Shazam are all small tools we use regularly – but all of these have AI behind them.
What are their equivalents in horticulture? Usually they fall under answering one of these four questions:
- What happened?
- Why did it happen?
- What is going to happen?
- What is the best thing that could happen?
So, to give a few examples: better data visualisation, automated diagnostics, yield prediction, and optimised control. A lot of these are big projects, but many are small tools that should still help make things easier for growers.
Projects like these are being done by Phil’s data science team, which currently consists of my supervisor Matthias, an experienced data scientist from Cool Blue; and Felipe, one of my biosystems engineering colleagues from Wageningen. Felipe recently finished his internship project and has joined the team as a full-time employee.
What’s my project about?
Modern greenhouses are full of sensors. Because of this, a grower can spend at least an hour going through graphs to find anything worth looking at. My project’s goal is to help automate this and help the grower know where to look.
That’s all I can say for now – partially because the project isn’t over yet, and partially because it’s full of secrets that will help Phil take over data-driven horticulture 😉 . But if you are a grower interested in what Phil is doing, feel free to get in touch.
It’s been great working in such a fun team with all sorts of backgrounds, from decades of horticultural experience to design to data science. Working at a company has been quite a new experience for me, especially in 2020. I’m grateful to be one of the few on the dynamic speedboat – whilst also being able to experience everything that comes with its mothership.
