Asimov’s worlds become real: robots and autonomous systems will provide us with food

The book steel vaults (1954) by Isaac Asimov develops its story in a future where humans flee to cities and robots do all the farm work. We are not confined yet and it seems that our level of agoraphobia is far from those highs. But the time is not far off when we have all the technological elements for fully automated agricultural work. A tomato or lettuce will reach our table untouched by a human hand at any point in the process.

The reduction of the rural population, added to the increase in general aging (most accentuated in rural areas), indicates that there will not be enough people in these areas to grow the food that we will need in ever greater quantities. The alternative goes hand in hand with the automation of agriculture, run by fewer people. In the same way that factories have been automated for the past 50 years, it is time for robots, as in Asimov’s novel, to take over the fields.

Autonomous agriculture: beyond the tractor

It’s not just tractors, autonomous agriculture uses technology to operate without human intervention in multiple processes: the decision of what to plant, where to plant it, when to do it, how to grow it, when to harvest it, etc.

Let’s take an example: let’s imagine a system that analyzes the prices of tomatoes on the market, monitoring the evolution of this product in the various forums where this information is offered. Through sensors in the field and remote sensing (satellites and drones), the system is able to evaluate the degree of maturation of the cultivated products.

In parallel, he analyzed food health regulations (written in human language) to determine whether they are fit for consumption after treating them for an infection a few weeks ago. On the other hand, the system has access to weather forecasting services to know the evolution of temperatures and rainfall. Relying on the digital twins that it has of the plants and plots, it simulates the evolution of the state of the product and makes the decision of when is the most optimal moment (maximization of agricultural yields) to instruct, now yes, the robotic systems that will proceed with the harvest and distribution through the most convenient logistics system.

Chimera or reality?

We already have scientific and technological foundations on which to build the realization of autonomous agriculture. The degree of development of each of the “pieces” is not the same. Autonomous tractors, yes, but also automated cultivation of products with a short cycle and a homogeneous harvest, such as lettuce; pepper-picking robots relying on computer vision for maturity recognition, albeit at speeds too slow to be profitable for now; crop or input control systems applied through precision maps; quads that travel the land at high speed taking hundreds (or thousands) of images of the soil in order to be able to determine, through suitably trained neural networks, the weed problems we have and what are the best treatments for them; or artificial intelligence that monitors agri-food prices and predicts their evolution.

The road that remains to be traveled is still long and complex. We need the integration of all progress into a common framework towards which efforts should be directed.

Shutterstock / MONOPOLY919

Problems we need to solve:

There are many fields of R+D+i that flow into autonomous agriculture. Without pretending to be exhaustive, let’s look at a few that might not be so obvious:

  • Process natural language. Rules, regulations and laws are written in a language for humans. We need technology to make the development of regulation (made by man) compatible with the execution of agricultural activities (made by machines).
  • Elements of the model of reality. Digital twins as a computer representation of the physical reality that is part of autonomous agriculture: the plants themselves, the farms and even the machinery involved. In them we can evaluate their evolution and simulate actions before putting them into practice.
  • Find the necessary and quality information that we need. The Web can offer us a lot of information, but it is full of poorly described resources, inconsistent information, gaps and data beyond the reach of browsers (“Deep Web”). Finding and determining its quality and reliability is a challenge for R+D+i.
  • Integrate data of different nature and granularity. You need to make data coexist together regardless of its origin, scale, data type, etc. We find resources from open data, geographic information arrangements, or culled from sources published on the web. But everyone must come together to answer the problems that arise.
  • Legislation, liability, insurance, society. Whose fault is it if an autonomous system misapplies a plant protection product? The farmer, the technology provider, the legislator? What new insurance models will we need? Is our society willing to eat tomatoes grown without human intervention? How will we educate her to this new reality? How to get him to trust her? Not only technology marks the challenges we face.

If we obtain collaboration between the agents involved (researchers, companies, public administrations, the political class and society), perhaps it will not be necessary to wait for the 47th century (the era of Asimov’s novel) to normalize autonomous agriculture. If robots take over the fields entirely, let them do it under effective human control. It is the moment.The conversation

To learn more about what is happening in the world of science and culture, join our Cultívate community, El Mostrador newsletter on these topics. sign up for free HERE

Follow us on

The Google News office

Leave a Comment

Your email address will not be published. Required fields are marked *