Abstract

Autonomous Greenhouse Challenge Project illustrates a case on AI and IoT-enabled cherry tomato production at a greenhouse enabled by Seeed’s SenseCAP LoRaWAN sensors and gateway.

Partners

Seeed Studio, Tencent, WAGENINGEN UNIVERSITY & RESEARCH, AiCU

As there has been a continuous decline in arable lands in the past decades due to human-made disasters and climate change, it is utmostly important for us to produce food more sustainably and efficiently. One of such methods is growing food inside greenhouses, as quoted by WUR: “Production volumes in greenhouses are typically up to 10 times higher than in the open field. At the same time, greenhouse production uses much less water compared to open fields.”

Similarly, the UN recognized the potential of greenhouse farming for generating stable income for farmers in the countryside, the World Economic Forum (WEF) also shared increasing urban farming trends amid the COVID-19. Indeed, we, as individuals, organizations, and enterprises, need to be prepared for the changing world affected by various force-majeure situations. By making modern techniques and systems such as AI and IoT more accessible, one of the most traditional industries like agriculture, can be transformed. Whether it is outdoor farming, tea growing on high mountains, fruit and vegetable cultivation, or any other categories of husbandry, technology can play a role in increasing efficiency, decreasing resource inputs, and incorporating sustainability to protect our environment, people, and the planet.

What’s the Project About?

In 2020, AiCU, one of our partners (Figure 2), won 2nd place at an event called “Autonomous Greenhouse Challenge September 2019 – June 2020” in Bleiswijk, the Netherlands. This event was aimed at inviting computer scientists and horticulture experts to challenge themselves, cutting-edge technologies, and manual greenhouse production, towards making a fully autonomous greenhouse operation possible. The event was jointly organized by Wageningen University & Research (WUR) and Tencent, with the goal of finding the best intelligent planting solution of cherry tomatoes, remote control of greenhouse crop production, and the best net profit optimization strategies in an autonomous greenhouse via deploying AI algorithms and sensors.

WUR is an academic partnership between Wageningen University and the Wageningen Research Foundation in the Netherlands. It is dedicated to healthy food and sustainable living environment research for government agencies and the business sector. Through incorporating academic research, real-life applications, innovative ideas, and diverse technology deployment, WUR creates scientific breakthroughs in the food and agriculture industry. WUR’s core values include society and wellbeing, natural resources and living environment, and bio-based food production, while emphasizing sustainability and corporate social responsibility (CSR) values.

Tencent is a world-leading Chinese IT company, specializing in innovative product development and services to upgrade people’s quality of living globally through their communication APPs, video games, high-quality digital contents, FinTech, cloud computing, and so on. Ever since its foundation in 1998, Tencent has been dedicated to using its “technology for good”.

Since 2018, WUR and Tencent have been partnering to organize challenge-based events to inspire the adoption of AI and IoT in cultivating food produce with less inputs of resources. In 2019, our IIoT product, the SenseCAP series, has been deployed by AiCU Team to monitor environmental data within the greenhouse, where cherry tomatoes are cultivated during the 6 months-long period of the Autonomous Greenhouse Challenge Project (Figure 3).

During the Autonomous Greenhouse International Challenge Project, WUR provided all the 5 teams with the same facilities and equipment, including greenhouse compartments, cherry tomato “Axiany” seedlings, substrates, lightings, advanced greenhouse simulators, some basic sensors, and the like. On the basis of these fundamental gadgets, each team was able to add extra sensors and cameras as needed. For this Project, data collected by the sensors are crucial to reflect the inner condition of the greenhouse, and precise control of the parameters can determine the growing status of the plants. By combining data with their own ICTs and machine learning algorithms, each team independently made decisions on the environmental settings of the greenhouse, such as lighting, temperature, humidity, CO2 concentration, nutrients, irrigation patterns, ventilation periods, etc. All the decisions made by the teams will determine their inputs and outputs, which would ultimately determine who would win the Challenge, based on what the Challenge Coordinator, Dr. Silke Hemming, describes as the 5 standards: “high quality, high yield, low-energy consumption, automation, and technology transferability” (Figure 4).