AIFARMS

AIFARMS Researchers Meet with Policymakers and Campus Leaders to Showcase Project Highlights

In January 2025, leaders from the Center for Digital Agriculture, AIFARMS Institute, and the University of Illinois Urbana-Champaign met with U.S. Representative Nikki Budzinski (IL-13), IL Senator Paul Faraci, and the ITI Institute to showcase how artificial intelligence is transforming agriculture for the better. From leveraging agbots to plant cover crops, contain weeds, and manage livestock, to developing helpful AI chat tools like CropWizard to answer crop-related queries via text and images. These advancements aim to transform the way we feed a growing population and make farming safer, more cost-effective, and more efficient.

“Things are changing…how farmers are farming, how they’re using conservation practices. We just saw a tool created here at AIFARMS that would help incentivize farmers to use conservation, cover crops in particular; that’s something we’ve been trying to encourage farmers to do,” said Budzinski. 

By showcasing the exciting research, innovations, and projects led by CDA and AIFARMS at Illinois, we collectively demonstrate the benefits and value of current and future investments in advancing U.S. digital agriculture. In this article, we highlight three innovations that represent the exciting work taking place at CDA and AIFARMS.

CropWizard

CropWizard is an interactive question-answering and decision-support service powered by generative AI for agricultural professionals in the United States. It is based on augmenting large language models, such as GPT-4 from OpenAI, by consulting an extensive knowledge base of agricultural technical documents. CropWizard is a multimodal system that accepts text and image prompts and provides textual responses. It can be accessed via a public web page or a software interface (API) over the Internet. The CropWizard system can deliver significant value to the farmer community and broad segments of the AgTech industry.

In this project, we are exploring ways to advance Generative AI capabilities (multimodal analysis, reasoning, planning, data discovery, and invoking relevant computational tools) to enhance CropWizard’s usefulness for commercial products and services. With its many applications, we are pursuing funding sources for a self-sustaining, long-term effort. Chat with the wizard here: https://uiuc.chat/cropwizard-1.5/chat

Ag-Bots for Cover Crops

EarthSense continues to work with AIFARMS team members on scaling up and deploying autonomous robots for cover crop planting. This effort is on track to be scaled up to 10,000 commercial acres for prototype evaluation and product refinement. The recent AIFARMS results from Thrust 1 this year on the CropFollow++ autonomous navigation and recovery algorithm were directly motivated by experiences at EarthSense, showing limitations of the previous CropFollow algorithm, and the new algorithm will be evaluated for deployment of commercial robots, potentially benefiting multiple robot models.

This project aims to increase the capabilities of mobile agricultural robots, focusing on smaller systems that can be deployed at scale across the field. We have been developing tools to improve navigation in highly uncertain environments and to improve the system’s robustness to uncertainty and failures. This year, we developed frameworks that enhanced the capabilities of the Ag-bots while making them more usable.

Horseradish Weeding Robot

The subproject aims to develop an AI-based solution for weed management in horseradish production farms in Illinois. We are focused on understanding the specific needs of horseradish production and creating a modular attachment for the Amiga robotic platform. This attachment will include a bottom-facing camera that continuously streams data to an AI model. The model will classify weeds and horseradish, and this information will be used to actuate a mechanical implement that autonomously removes weeds through either hoeing or pulling.

Our approach involves conducting multiple data collection campaigns across various field conditions and growth stages, both at IAF Farm and commercial growers’ fields. This will help train robust AI classifiers for effective weed and crop classification. Collaboration with farmers is essential for understanding, refining, and deploying these solutions to promote technology adoption.

Over the past year, we modified the camera arrangement on the robotic platform to capture top-view images of crops and weeds. We collected data on horseradish and weeds from both the IAF Farm and a commercial field in Collinsville, IL. The CDA REU intern was key in setting up the camera and collecting field data. The intern trained a basic YOLO v8 object detection model, achieving a decent accuracy (mAP of 0.79). However, this model was based on manually collected data using a camera on a monopod (see left image). Moving forward, we will use the robotic platform to collect data frequently and further train AI models. We have also collected drone images of the commercial field to see if we could develop a weed density map. A GRA hired through this subproject will process these datasets and focus on training and tuning these models to improve inference time. Our ultimate goal is to develop modular attachments that can effectively ‘detect and destroy’ weeds, transitioning research outcomes into practical, market-ready solutions for horseradish farmers.

Check out this article by the College of ACES to learn more about the visit, and/or watch a news clip about the coverage of the AI Agriculture Technology Farm Bill below.

The Artificial Intelligence for Future Agricultural Resilience, Management, and Sustainability (AIFARMS) Institute is supported by the USDA National Institute of Food and Agriculture and the National Science Foundation.

©2021 Board of Trustees of the 
University of Illinois  |  Web Privacy Notice

ABOUT COOKIES

Close

About Cookies

Cookies and related technologies (herein “Cookies”) are small text files that a website saves on your computer when you visit the site. Cookies the University sets are called first-party Cookies. The data collected might be about you, your device, your preferences, or your login information. This data is mostly used to make the website work as expected so, for example, you don’t have to keep re-entering your credentials whenever you come back to the site. Cookies set by third parties are called third-party Cookies. We use third-party Cookies for analyzing website traffic and our advertising and marketing efforts. We have divided the Cookies we use into the following categories: Strictly Necessary, Performance, Functional, and Targeting. Under each category heading below you will find a general description of the Cookies in each category. You can change your browser settings to block, delete, or alert you to Cookies. The Help menu on the menu bar of most browsers will tell you how to do that. However, if you do, you may have to manually adjust preferences every time you visit a site and some features may not work as intended.

Read More…

Cookie Categories

Strictly Necessary Cookies are first-party Cookies that are necessary for the website to function. They can be either permanent or temporary and are usually only set in response to actions made directly by you that amount to a request for services, such as logging in or filling in forms. For example, we use Strictly Necessary Cookies to handle user registration and login. Some sites require the use of Strictly Necessary Cookies to access the site, such as University websites requiring University credentialed authentication. If you set your browser to block or delete Cookies, you may not be able to access the site or some parts of the site will not work.

Always Active

Performance Cookies allow us to count visits and traffic sources so we can measure and improve the performance and effectiveness of University websites. Performance Cookies also help the University understand which webpages are the most and least popular, see how visitors move around the site, and determine whether webpage content is relevant to user interests. Performance Cookies may be first-party or third party, permanent or temporary, and do not personally identify individual visitors. Some Performance Cookies are “analytics” Cookies (e.g., Google Analytics), using third-party software tools, which help us understand more about how our websites are used and where visitors come from by collecting and aggregating anonymous information on the pages visited and any advertisements viewed. The University does not take responsibility for the collection, use, and management of data by any third-party software tool provider unless required to do so by applicable law. If you set your browser to block or delete Cookies, some site services and functionalities may not work.

Always Active

Functional Cookies enhance the performance and functionality of our websites but are non-essential to their use. These permanent Cookies allow our website to remember information from your previous visits, such as details you submitted before or your previously stated preferences. These Cookies may also be used to provide services you request, such as newsletters or publications. They may be first- or third-party Cookies that enable services we have added to our webpages. If you set your browser to block or delete Cookies, some or all of these services may not function properly.

Always Active

Targeting Cookies are used to deliver content tailored to your interests and may be temporary or permanent. They may also be first-party or third-party Cookies. Targeting Cookies are based on uniquely identifying your browser and device; they do not store information such as your name. The University may use targeting Cookies prepared by the University, its third-party contractors, or advertising partners to provide you with personalized University display advertising and promotional material about the University and its programs. The University may also allow third parties to place Cookies on your device that collect and use anonymous information about your visits to, and interactions with, our websites to personalize advertisements and promotional materials for University goods and services. Targeting Cookies may be used by our third-party contractors or our advertising partners to build a profile of your interests and show you relevant advertisements on other sites. We may share information about your use of our site with our social media, advertising, and analytics partners who may combine it with other information that you have provided to them or that they have collected from your use of their services. If you set your browser to block or delete Cookies, you will still see advertisements, but they will be less targeted to your interests.

Always Active