Tools that offer early and accurate insight into plant health鈥攁nd allow individual plant interventions鈥攁re key to increasing crop yields as environmental pressures increasingly impact horticulture and agriculture.

In response to this challenge, Cornell researchers have developed a soft robotic device that gently grips and injects living plant leaves with sensors that help it detect and communicate with its environment. The robot can also inject genetic material that could be used for bioengineering plants in the future.

The device allows for safe, repeatable delivery of sensors and genetic material in a reliable, plant-safe way鈥攁n essential step in precision, data-driven agriculture. The team's findings are in Science Robotics.

"Plants, like people, have different responses to the environment, and precision agriculture is an effort to move closer and closer to single-plant-level intervention鈥攁nd the soil surrounding it," said the paper's senior author, Robert F. Shepherd, professor in the Sibley School of Mechanical and Aerospace Engineering at Cornell Engineering, and a research lead at CROPPS.

Horticulturists, farmers and agriculturalists face rising pressures from environmental impacts, such as drought and fertilizer runoff. By implanting sensors into leaves, researchers can monitor the impact of drought or an overdose of fertilizer on the plant.

To demonstrate, the team used the gripper to deliver two types of probes. The first, AquaDust, is a tiny gel particle that fluoresces in response to water stress, allowing researchers to noninvasively monitor a plant's hydration levels. The second probe, RUBY, is a gene-encoded biological reporter that causes red pigmentation to appear where genetic transformation occurs within the plant.

"AquaDust allowed us to 'see' the water stress inside a leaf, and similarly, by injecting a bacterium that transforms the injection region with RUBY reporter genes, we were able to 'see' that this part of the leaf experienced a genetic transformation," said first author Mehmet Mert Ilman, previously a postdoctoral researcher in the Organic Robotics Lab and now an assistant professor in mechanical engineering at Manisa Celal Bayar University in Turkey.

"It was fascinating to be able to robotically transform the local genetics of the plant leaf and then see it change back," he said.

The researchers tested the device on sunflower and cotton leaves鈥攑lants known for their structural resistance to infiltration. The gripper achieved more than 91% success in delivery while causing significantly less damage than syringe-based methods and expanding the effective infiltration area by more than 12 times.

The soft robotic system works hands-free, delivers materials more evenly and causes little to no damage, even in tough and durable species like cotton. The technique is an improvement over traditional manual methods such as vacuum infiltration, which uses low air pressure to force liquids into plant tissues, or needle injections, which can injure leaves, are labor-intensive and often fail in tough plant types. This is especially important for horticultural crops鈥攕oft-skinned plants cultivated for their fruits, vegetables, flowers or ornamental value.

The device applies gentle, uniform pressure through a sponge tip that holds the nanoparticles or genetic probes. The design of the soft material and actuator (the portion of a machine that produces force or torque) were optimized through simulation software and 3D printing, allowing the gripper to function with a variety of leaf types and shapes.

"Its low stiffness allows it to warp the gripper's shape to adapt to the orientation and surface of the leaf with few health implications to the leaf," Shepherd said. "The shape of the extending actuator allowed for a large displacement and the ability to adjust orientation without bulky or complex motor control."

The research lays the foundation for real-time, minimally invasive plant monitoring, Shepherd said.

"Soft grippers to inject physical or biological probes unlock new and incredible capabilities," he said. "The immediate use of our system would likely be in greenhouses, where a robot would persistently inject and monitor individual plants to infer how much water they need."

In the long term, similar grippers could be used to deliver or retrieve other diagnostic materials, including sensors for nitrogen uptake, disease presence or even real-time metabolic changes, opening up new possibilities for smart agriculture and plant research.

"New nanoparticles will also eventually be created that will inform us about many other health aspects of the plant," Shepherd said. "With this information, plants will yield more, and we will waste less."

The team is now exploring the integration of the gripper into robotic arms for automated greenhouse systems, with the long-term goal of adapting it for field-deployable platforms.

"Once translated out of greenhouses, the implications will be larger," Shepherd said. "I am particularly interested in limiting the waste streams into lakes to prevent harmful algal blooms."

Co-authors include researchers from the Boyce Thompson Institute, the Smith School of Chemical and Biomolecular Engineering and the School of Integrative Plant Science in the College of Agriculture and Life Sciences.