Let me share with you why I love agriculture and why I think that as engineers, we have the possibility to change the world using technology. I come from a family that has devoted much time to cultivating the land and grazing the herd for years. My ancestors always repeated the same patterns year after year and only made variations depending on the behaviour of the weather, but what happened? Because my ancestors were late in time/acted late, and I always heard the same sentence: “We cannot do anything against the climate; we have already done everything we can.” Similarly, I have always been interested in technology and how it can help improve the world. When I was a kid I would take apart all the tech stuff I could to understand what was inside and find clever ways to use it for other purposes. That’s why as an adult, I asked myself the question:
What are the challenges of modern agriculture?
• Enable the healthy growth of plants and make them more productive
• Achieve better productivity levels while maintaining safe working conditions for farmers
• Improve the sustainability of the fields
• Keep agriculture economically viable, to prevent farmers from abandoning their land
How can we help?
Currently the costs of new technologies have been reduced to levels that make them accessible to most of the population. IoT, Cloud and Mobile technologies are key in the transformation that modern agriculture requires. However, one of the main challenges for the introduction of technology in rural areas is connectivity. Until a few years ago, the coverage was not very good. The adoption of 4G and 5G by telephone companies as well as the emergence of new technologies/protocols such as LoRaWAN, ZigBee, Z-wave, NB-IOT, etc. now offer a suitable level of connectivity to transmit data in real time at a very low cost. One way to use technology in agriculture is to create solutions that can measure the behaviour of the earth in real time. Here IoT plays a particularly key role, measuring the humidity, temperature and conductivity at the roots of plants, as well as the temperature, humidity, pressure and rainfall at ground level. Current IoT systems allow us to collect this data in real time, but this leads us to another challenge: Do we have to store all the data?
Make this simple calculation:
We take measurements from six different sensors (humidity, conductivity, temperature …) every 15 minutes for 365 days a year. Six sensors/ha x 4 measurements per hour x 24h x 365 days = 210 240 data records/ha.
According to the FAO, currently 11% of the Earth’s land (13.4 billion ha) is cultivable, so we have 1.474 billion ha times 210 240 measurements/ha per year, making a total of 309.894 trillion.
Almost 310 trillion recordings per year (20 billion Tb/year (2Kb per measure). Could you imagine what kind of databases we would have to use or how much energy we would spend?
Obviously, the answer is: No. We need to aggregate data and analyze when it has occurred. This is where technologies such as big data, elastic search, kibana, hadoop, distributed cloud, hybrid clouds, etc. play a key role.
With all this raw and processed data, we can extrapolate using meteorological services to generate data using predictive models with Machine Learning (AWS Machine learning, Google AI, Azure ML) environments to help farmers make reasonable decisions at key moments.
To support the mass collection and processing of data, the following technologies are used: Message Brokers such as Kafka, RabbitMQ, Azure Service Bus, Google PubSub Cloud, Amazon MQ, etc.Once the data and generated models have been processed, these data must be made accessible to the user. At this point, we face another challenge: Which technology to choose? SPA, Web, Hybrid Application, Native Application, Cross-Platform Application, Angular …
Because the data that the farmer must receive on the mobile must be very specific and does not require the high-performance capabilities of mobile phones, a good option is to choose Cross-Platform development for the mobile (like Xamarin, React Native, Ionic, Flutter, etc.) and a website for computer access with more detailed information and expanded options (like Angular, React, etc.)
According to the FAO, 70% of world water expenditure is from agriculture; 50% of this expenditure can be reduced by IoT and Machine Learning systems. Moreover, by providing accurate information on the state of the land and the need for farmers to plant in real time, we can reduce fertiliser and pesticide costs between 40 to 60% and increase production between 30 to 40% based on my current experience developing these types of solutions.
We will also help to reduce the contamination of soils and water reserves since we will be consuming only the amount of fertilizer and water that crops/trees require at the times that they require it. All of this will help us design better preventive treatments that are safer than the current corrective treatments, which have a huge impact and a greater residuality on fruits. Let’s do a quick calculation again:
Fruit field (1ha):
Treatment | Quantity | Cost | Total Euro/ha per month | IoT savings | Total + IoT savings Euro/ha per month |
Water | 160 m3/ha per month | 0.12 Euro/m3 | 19.2 | 50% | 9.6 |
Fertilisers | 100 kg/ha per month | 1.60 Euro/kg | 160 | 50% | 80 |
Phytosanitary | 4 treatments/ha per month | 90 Euro/treatment | 360 | 50% | 180 |
Total costs | 539.2 | 269.62 | |||
Total production | Price | Total | |||
Without IoT | 20 000 kg/ha | 0.90 Euro/kg | 18 000 Euro | ||
With IoT | 30 000 kg/ha | 0.90 Euro/kg | 27 000 Euro |
Conclusion:
“We aim to make IoT technology understandable for people who spend most of their time working on the farm and are normally outside getting their hands dirty. Once we engage them and help them understand that there is nothing to fear, then they’re in.”