Netherlands-based Wageningen University is holding the third edition of its Autonomous Greenhouse Challenge, where international teams compete to grow crops in greenhouses using AI and automation.
A first and second edition of the challenge grew cucumbers and tomatoes, respectively. For both of those crops, a level of autonomy was involved, though human intervention was still required during the grow process. The third edition, where teams will grow lettuce, differs in that participants must figure out how to grow their crops from seed using a fully automated algorithm.
Wageningen said in a post that the first and second editions of the challenge have shown “that artificial intelligence can potentially be superior to human intelligence, hence can potentially control indoor farming in the future. The goal of the third challenge will be fully automated control.”
Wageningen is one of Europe’s most well-known innovation centers when it comes to food, food tech, and food innovation. Scientists and researchers from here work on everything from alternative proteins to biodiversity initiatives to gene-editing technologies and, of course, greenhouse innovation.
One of the goals of the Autonomous Greenhouse Challenge is to “connect the world of artificial intelligence (AI) and food production, create more knowledge, make this knowledge publicly available and thus contribute to the worldwide efforts of making our food systems more sustainable.”
Part 1 of the competition is an online challenge meant to test participants’ machine learning and computer vision skills, as well as attract talent from the realm of AI into the horticulture space. Part 1 is currently underway and will wrap on July 14 when a winner is chosen. Following that will be a 24-hour hackathon, which anyone can join, even if they did not participate in Part 1. Five teams will be selected from this session to go on to compete in Part 3, the actual greenhouse challenge.
For the latter, each participating team will be assigned one compartment inside Wageningen’s high-tech greenhouse facility. All compartments are identical and have different actuators that control the grow conditions inside: heating and shading systems, lighting, water input, etc. Compartments are also equipped with sensors to measure temperature, CO2 levels, PAR light, pH and the electrical conductivity of fertigation water. Data from these sensors can be used to inform the algorithms to decide on the compartment’s control settings.
Teams will be allowed to do a test cycle with a first crop before they must transition to growing the lettuce using fully autonomous algorithms “make choices with respect to the control settings, to remotely control crop growth.”
Even without automation in play, lettuce is typically a more “hands off” crop than tomatoes or other vegetables. The competition seems not so much about automating indoor farm labor as it is about showing the public the benefits automation can bring in terms of growing plants faster and at higher yields without sacrificing quality. It will also produce more data, which experts have long agreed indoor farming needs more of.
In the case of Wageningen’s challenge, the public will be able to follow crops’ growth along in real time via a website, and the datasets will eventually be released to the public. The winner of the competition will be the team with “the highest net profit in the end,” according to the challenge rules.
