High-Tech Meets High Stakes: AI-Powered Field Robots Revolutionize Crop Breeding

We’re thrilled to share that a major collaborative research project carried out Corteva Agriscience, University of Illinois Urbana-Champaign, and EarthSense has been published in Nature Communications Biology. This peer-reviewed publication confirms what we at EarthSense has been working toward: high-throughput, autonomous field phenotyping is no longer a dream — it’s a reality.

You can read the full paper here.

Over five years and across 142 research fields, the Corteva team collected trait data from nearly 200,000 maize plots — using our TerraSentia platform to measure traits that really matter. Crop breeders at Corteva and UIUC used TerraSentia robots to repeatedly collect high-resolution, in-canopy data from maize breeding and research fields. The study reports on results of four of the key traits: Leaf Area Index, Plant Height, Stem Width, and Ear Height.

These traits unlock critical insight into:

  • Drought tolerance

  • Fertilizer use efficiency

  • Standability and biomass accumulation

  • Yield potential

Because we measured traits multiple times during the growing season, we enabled researchers to analyze how hybrids respond to the same environment over time — revealing a powerful new understanding of genotype-by-environment-by-management (GxExM) interactions. This is exactly the kind of insight that is essential to power the next generation of breeding: faster, smarter, and more resilient crops.

Despite the leaps in genomic sequencing, one key bottleneck has remained: understanding how different genomes perform in real-world field conditions. Traditional field phenotyping is expensive, slow, and extremely limited in scale. The TerraSentia system breaks that bottleneck! The AI-driven robots and ML based analytics algorithms that constitute the TerraSentia system deliver accurate, repeatable, and affordable phenotypic trait measurements from multi-sensor data — on a scale that was previously unthinkable.

Jason DeBruin, a trait evaluation scientist at Corteva Agriscience, added: “With TerraSentia, we’re able to measure traits that were previously out of reach due to labor constraints. This increases our knowledge of germplasm and traits, potentially accelerating our pipeline and improving our ability to develop the next generation of crops.”

“With five granted patents and more in progress, and a powerful, scalable data pipeline, we’re ready to bring high-throughput phenotyping to farms and research stations across the world. This is just the beginning,” said EarthSense CTO, Girish Chowdhary. “We’re scaling precision agriculture globally — helping crop breeders make better crop varieties, faster, for a rapidly changing planet.”

“This paper clearly shows that crop breeders can now consistently and repeatedly measure essential, but labor-intensive, crop traits at commercially meaningful scales — without breaking the bank or burning out their field teams,” said EarthSense CEO, Chinmay Soman. “Our solution is great way for leading crop breeders to gain and expand competitive advantage.“

This paper marks a high point in an extraordinary collaboration:

  • At Corteva Agriscience, Jason DeBruin, Haley Underwood, Rebecca Hensley, Sara Tirado Tolosa, and their colleagues deployed fleets of TerraSentia robots across the U.S. and overseas.

  • At University of Illinois Urbana-Champaign, Steve Moose and Grace Nystrom led the charge on exploring the scientific frontiers of phenotyping—with support from the College of ACES, the Grainger College of Engineering, and the Center for Digital Agriculture.

  • At EarthSense, CTO Girish Chowdhary and our global, multi-disciplinary team of AI and field robotics experts built and deployed the robotics and AI technologies that made this possible.

We also owe tremendous gratitude to:

  • The ARPA-E TERRA program, led by Joe Cornelius, David Lee, and Krishna Doraiswamy, which sparked this vision and funded the initial development of the core technologies.

  • The TERRA-MEPP project at UIUC’s Carl R. Woese Institute for Genomic Biology, led by Stephen Long, Carl Bernacchi, and Don Ort.

  • The National Science Foundation’s SBIR program, with critical support from Anna Brady-Estevez, which helped translate our ideas into scalable products.

Corteva’s large-scale, multi-year deployment proves that TerraSentia isn’t just for pilots — it’s a production-ready platform for real breeding programs.

If your program still relies on manual measurements or one-off drone snapshots, it’s time to ask: Why wait? TerraSentia is ready now — and the advantage it delivers is only growing.

Want to modernize your phenotyping workflow? Reach out to us at EarthSense to get started with TerraSentia today!

in-the-news, partnershipEric Kurniawan