Wednesday, 3 March 2021

Using AI & Machine Learning to Improve Local Farming: SELA’s Food & Agritech Project 2020-21

Natasha Anderson & Sam Casadei, SELA Cohort 2019

This year’s integrative project theme is artificial intelligence and machine learning. It provides us with the chance to work alongside industry representatives, and local companies to promote the uptake of digital solutions, whilst also challenging us to apply the leadership skills that we have been developing throughout our time in SELA. Our team hopes by taking the lead in the small-scale farming sector, we can enable new forms of digital food systems data collection and can engage the Sheffield digital industries sector and the Sheffield food & drink sector in a range of innovative data and digital collaboration research activities. 
The Food and Agritech Project aims to find a way to collect data in small farms so AI and Machine Learning can be used to increase productivity. Working together with Sheffield-based food cooperative Regather, we are hoping to do this through prototyping and collecting data on basil growth as it’s easy to grow at home. We will collect data using Raspberry Pi technology combined with a range of sensors, capturing data on light, temperature, humidity and soil moisture.

We have designed a prototype consisting of a propagator containing basil plants and an LED strip as a light source. This aims to recreate a polytunnel, which is a type of greenhouse used in small-scale farming. We will use the inexpensive microprocessors Raspberry Pi A+ and GrowHat, and sensors from the local company Pimoroni. This kit will detect all the key inputs needed for successful growth. As we will be coding the system, this will be a great opportunity to learn more about Raspberry Pi, as well as apply our academic learning. This setup has been great during the COVID-19 pandemic as it allows us to work on the project at home, without needing access to laboratories or specialist equipment.

Using this setup we can collect data and see how this has affected the growth of the basil and we can see how it reacts to different conditions and use this data to formulate a ‘best case scenario’ set of conditions through AI and machine learning.

We are hoping that at the end of the project we can reach number of targets:

1. Make personal home grown produce more attainable to the ‘average’ consumer

Starting to grow produce at home can often seem very daunting to those that have no experience, wondering are they watering enough, or what temperature the plant should ideally be at.

This project aims to create equipment that is ‘plug and play’, with an easy to use guide, so that even those with no prior experience of Raspberry Pi and local food growing will be able to know exactly what conditions their plants need.

This research is currently centered on generating data from basil as a test crop, and we plan to work with Regather to continue the experiments with other crops, as a community engagement event. Ideally, the user will select a type of plant to grow, and the monitoring system will be able to give insightful prompts on the conditions of the plants, allowing for easy care with best results.

2. Create an interest in STEM for young children, with an easy to use, safe, plug and play kit.

Practical hands on kits have proven invaluable for children as a result of the COVID-19 Pandemic [1]. In a similar vein, a ‘do it yourself at home’ kit is likely to have a similar uptake. Showing how user-friendly the kit is will be another example of how technology can be beneficial, and aims to get children hands on and creative. This also promotes the benefits of practical based learning. Although outside the immediate scope of the project, it would be feasible to roll out the system to a wider audience, and work with local schools to look to integrate the kits within a classroom environment, to act as a central source of interactive teaching.

3. Drive interest from the younger generation into how food is produced

With the continuing emphasis on healthy eating and food education, this kit will supplement the existing education available. The growing renewable movement, and focus on reducing carbon footprint, leads in well to home grow produce, and starting with an easy to care for herb is just the first stepping stone to promoting growth of other vegetables. It also emphasises the importance of understanding where food comes from, the conscientious nature of choosing what food to buy, and the ethical choices that come with that.

4. Collaboration with local companies, and local support groups, to drive nationwide interest in the project.

One of the fundamental aspects of this design was to work closely with local suppliers in Sheffield, in order to promote locally available material. Pimoroni has been the main supplier for the technology for the project, and also has a highly active forum of nationwide users of Raspberry Pi. They were also able to provide advice on the best products to use.

A local company, Grobotic Systems, has also been directly involved with shaping technology, and several meetings have been conducted in order to design a suitable sensing design. They have direct experience with growing herbs and vegetables in controlled environments, and their experience was crucial to ensure that this design would be optimised for its specific use.

In summary, our project aims to make growing your own food more accessible to the everyday user, firmly placing quality control in the hands of the end user. Through easily understandable data collection, education on the important factors in food production is made possible, and aims to shift to more local growing. This shift would increase biodiversity, public morale, whilst also decreasing the effects of large scale farming on the environment. Less Carbon Dioxide pollution from the transport and preparation of the produce is just one way home grown produce is beneficial for the environment.

[1] Kosmos, S., 2021. Surge in demand for STEM toys and science kits ‘will continue well into 2021’ says Thames & Kosmos - ToyNews. [online] ToyNews. Available at: <>

1 comment: