Autonomous Greenhouse Challenge – The Conclusion!

The Greenhouse Automation Challenge is a yearly competition in which multidisciplinary teams of experts and students in the fields of machine learning and horticulture are tasked with operating greenhouse production remotely in order to grow high-quality crops. This remote growth is achieved with the help of specialized sensors and specially tailored algorithms. 

The team’s main goals are to grow a high-quality cherry tomato crop within 6 months while maintaining high production yields and low resource usage. During the 6-month period, the teams were not allowed to enter the greenhouse units and had to make changes/manage the crops entirely remotely. Heliospectra sponsored the event by supplying ELIXIA lights which joined existing HPS lights in the greenhouse units to create a hybrid lighting system (HPS and Heliospectra’s ELIXIA LED Grow lights together). This way competitors gained higher control over their lighting and heating parameters and were able to fine-tune for suitable growth conditions through even the harsh winter months.

View our Webinar on Growing Tomatoes Autonomously >> 

Who Won? 

During the awards ceremony, the organizers made it a point to draw a lot of attention to the unique and valid approaches that each team took over the course of the six-month competition. The judges‘ panel illustrated this by highlighting things that each team had done right – and wrong. However, in the end, only one team was able to truly win, that team being the AuTomatoes, led by team captain Leonard Baart de la Faille. This result was established by a panel of judges judging their methodology and results against the following criteria: Net Profit, Sustainability, and AI Strategy. 

What Did The Overall Results Look Like? 

This year’s results seemed rather unpredictable and difficult to judge with respect to the quality/quantity of tomatoes grown. This was because while each team employed different strategies, they for the most part ended up with fairly similar results. Two teams specifically had similar or higher production than the other growers and were able to just outgrow the “typical farmer” which was used as a control variable for the competition. 
 
However, the really interesting results from the competition are based on the resource control and savings that each team was able to enjoy. Each team was able to precisely control temperature, lighting, CO2, and irrigation amongst other factors, and saw huge savings in these areas thanks to the AI (artificial intelligence) systems that they were using. As a result, their net profits increased in relation to the control farmers as they could now produce similar amounts of crops for less cost. 

What’s The Key Takeaway From All of This? 

The major implications of this competition are that growers could soon start using AI systems to control their growing parameters and save money and resources in doing so for automated greenhouses. The AI systems have been found to be more accurate and are able to outperform human growers in making these resource changes, this would thus allow for more eco-friendly production and cost savings. These efficiencies increase as the greenhouse size increases, making this an attractive model for larger installations. 

Another major implication is that because of the AI systems, the crops were grown remotely. This is especially important today due to the situation with COVID-19 where growers have had trouble working on their crops. AI systems make these problems essentially nonexistent, by allowing for full remote and autonomous control of greenhouses. Because of this and the increased efficiencies, these AI-driven systems are thought to be the growing method of the future!   

Interested in learning more about the competition and what the future of autonomous cultivation could look like? Then be sure to join us for our “By Growers For Growers” webinar where we’ll be speaking with the AuTomatoes team captain, Leonard Baart de la Faille, and the leading organizer behind the event Dr. Silke Hemming! Learn more.