Researchers to develop technology for improving berry monitoring, harvesting, shipment

The agricultural industry has long faced challenges to berry harvesting, from the delicate nature of picking the fruit off the plant, to labor shortages and the physical impact of this labor on workers. To investigate innovative methods to support the berry growing and harvesting process, UC Santa Cruz Professor of Electrical and Computer Engineering Ricardo Sanfelice is launching a new agricultural technology project.

This two-year project is a collaborative effort with Professor Stefano Carpin at UC Merced and Professor Manuel Navarro-Gutiérrez and researcher Jorge Isaac Chairez Oria at Tecnológico de Monterrey (Tec de Monterrey) in Mexico. The group received funding from Alianza MX, a University of California systemwide initiative to enable collaborative academic and research programs between scholars in UC and Mexican institutions.  

“Berry harvesting involves a lot of very interesting fundamental robotics control problems, which we have researched in different contexts and we can apply to this very important problem,” Sanfelice said. “Then there's also the societal impact aspect, which I'm very interested in: how do we actually support the workforce in order to create technological solutions that are not only efficient, but also ethical?”

Better berry picking

Berry picking is a very physically demanding job, and the researchers involved in this project aim to develop technology that can reduce harm to workers and aid in situations in which picking the berries could be particularly labor intensive. Their technological innovations will support berry monitoring in greenhouses, harvesting, and shipment across the world.

To monitor the health and maturity of berries grown in greenhouses in the U.S. and Mexico, the researchers will implement a sensing system with algorithms that use computer vision and machine learning to detect the fruit on the plant and identify ripeness or disease. This will inform when to pick berries or intervene for the plant’s health. Experts at Tec de Monterrey will take the lead on algorithms to catalog and detect berries.  

“It's certainly a challenge to determine when to pick the fruit, and to do so at a rate that is comparable to the human worker, the human decision maker,” Sanfelice said. “With a lack of potential workforce and the need for this fruit to be available, we are asking how we can automate some of these stages.” 

The researchers will also investigate methods to harvest berries with robotics or other technological solutions, such as vibrations that shake the berry off the plant. This is a complex challenge due to the need to harvest the fruit in a precise, careful, and efficient manner while all the fruit may not mature at the same time. Humidity and moisture in the greenhouse environment could also affect the harvesting process. 
Sanfelice’s Hybrid Systems Lab in particular will work to design vision-based algorithms that can execute the grasping motions needed for berry picking, applying the correct amount of pressure at the right places on the fruit.

“This requires coordination between different systems, for example combining vision with decision making and physical motion to create safe and careful handling of the fruit,” Sanfelice said.

“At UC Merced, we have built a diverse portfolio of robotics projects aimed at supporting sustainable agriculture,” Carpin said. “This new initiative takes us in an exciting new direction, as we had not previously worked on challenges related to berry cultivation. Specifically, we will focus on optimizing the routing of robots in orchards to enhance efficiency—an important and complex optimization problem with significant practical implications.”

After the berries are harvested, they must be handled and shipped, a process which sometimes takes more than a month when the fruit is being sent by boat across the world. The researchers will look into ways to prolong and maximize the quality of the berries.

Collaboration with industry

Throughout the process, the researchers will collaborate with partners in the agricultural industry, including Avoberrys el Valle to make sure they are addressing important problems for growers and workers. 

“We want to be informed about the real challenges here, versus solving some sort of academically interesting problem that might not have any relevance to industry needs,” Sanfelice said.

Sanfelice hopes that this collaboration will also allow students involved in the project to contribute to the mission of the industry partners, providing them with valuable training as they enter the workforce. The group plans to present findings at two forums: “Expo Agricola Jalisco,” an agricultural technology exposition in Mexico held yearly in May and “Colloquium on Discrete Event and Hybrid Systems” organized by the scientific community on discrete systems in the city of Guadalajara, Jalisco, Mexico.

The project’s funding includes opportunities for two students from Tec de Monterrey to visit UCSC and UC Merced, and for one student from each of the UC schools involved to visit Tec de Monterrey. Additionally, one student from UC Merced will travel to work with the group at UCSC.

UC Santa Cruz hosted a kickoff meeting for this project on October 4.