Robots in the Fields: How Solinftec is Cultivating Brazil's AI-Powered Agricultural Revolution

In an enlightening interview, Bruno Pavao, Chief Robotic Operations at Solinftec, offers a glimpse into the cutting-edge world of AI and robotics in Brazilian agriculture. Pavao discusses the practical applications of AI in farming, with a focus on Solinftec's groundbreaking autonomous robot, SOLIX. As Brazil faces unique agricultural challenges, including climate issues and the absence of government subsidies, Pavao explains how AI-driven solutions are revolutionizing farm management. From weed identification to yield prediction, he outlines how these technologies are helping Brazilian farmers increase productivity and sustainability. Pavao also shares insights on data privacy, future developments, and the potential long-term impact of AI on Brazilian agriculture.

Bruno Pavão

Chief Robotic Operations at Solinftec

To the best of your knowledge, what are the current practical applications of AI technology in Brazilian agriculture?

We are seeing at the farms A.I been applied focus in this points:

• Weed identification

• Identification of pest infestation

• Potential yield 

• Symptoms of stress and diseases

The use of robots and drones are the path to collect this data and the platforms that are been created to interact with the farmers are been improved everyday with new insights from the crop and researches tht are working with the sensors and plants.

Solix Ag Robotics was introduced as Solinftec’s first technology to use artificial intelligence (AI) to monitor and care for every part of farm plots. Can you tell us more about Solix? How does it work? 

SOLIX is the first Brazilian Autonomous Robot to interact as a commercial product with the farmers and cooperatives in Brazil, United States and Latin America.

The robot is designed to live in the field, 100% electric powered by battery and solar panels that can provide for the robot a self-feeding environment to work on the field everyday.

The robot is equipped with 12 cameras to use A.I recognition developed by SOLINFTEC to identify pest infestation, plantability, distance of plants, doubles, height of the crop and weed identification. For the weed identification SOLIX can act in real time to apply herbicides as a spot application to control the weeds and reduce the use of herbicides. The main goal to build a ROBOT as a platform is to be able to do all the scouting missions and spot spray applications at the same time.

Can you explain the process of training and updating your AI model to ensure accuracy and relevance as agricultural conditions and practices evolve over time?

The process to create our own tool of A.I started in 2020 with the first prototype developed to test the concept of the robot. The data and photos collected to identify all the goals to create insights and interpretation are collected every mission. We created a new department inside of the company to analyse and segment all the data that we use to create A.I trainings that ran inside of the brain of the robot. In average more than 2 millions of photos and data lines are analysed per month to increase accuracy and knowledge for the data base to be able to create a new weekly training model for the robots in all the crops that we are using the platform 

Solinftec has a solid base of clients and with the diversity of backgrounds (soil type, variety of plants and crops) is faster to adapt the SOLIX in other conditions. As the frist company to deliver to the farmers a tool that is fully autonomous, reducing the costs to spray herbicides and increase yield expectation is the most powerful relevance that the farmers are looking for

What measures have you taken to address potential concerns regarding data privacy, security, and bias in your AI systems, particularly when dealing with sensitive information from farmers and agricultural operations?

Solinftec is using the high level protocol to ensure that all the data and information collected in the field are secure. All the platform is controlled by user access and traceability for the farmer to understand where the data is and available to access any time they need.

We work with data from the farmers since 2007 with our base solution on telemetry and operational data over 10 millions of hectares without any concerns.

What new developments or breakthroughs is your company working towards to further enhance agricultural productivity, sustainability, and resilience?

We are working on the creation of a A.I Generative that is capable to understand and provide insights about the crop evaluation using all the informations collected by the robot and mixed with the operational data that is possible to collect. The kind of insights mixes climate data, planting data, soil condition, genetic variety used and others.

The development is running fast, and we will be able to test a new version in 2024 season in Brazil and this will be the answer to find more space for the use of A.I to increase yield and reduce costs.

What far-reaching impact will the application of AI technology have on agriculture in Brazil?

Brazil is a country where the farmers do not receive government subsidies for agricultural production, and they are extremely affected by the climate challenges and the delay on the development of new molecules or genetics that can face the challenges. By that we have a scenario that is running against time to break the cycle of resistance that we are facing using the conventional management strategies. 

The use of A.I as a path to create a new management strategy will impact positively the farmers when they are able to reduce cost and increase yield, because at the end of the season, Brazilian farmers don’t have the secure about their profits with a conventional strategy in high cost and average yield.

This story was initially published in the 2024 Latin America Focus. Download the magazine to read more stories.

Please contact Christina Xie at christina@agropages.com if you would like to share your company story, contribute articles or advertising with AgroPages.