Cornell University is leading a new $7.5 million national research effort aimed at developing robots that can handle some of the most labor-intensive jobs in fruit orchards, from pollination and thinning to harvesting and weed control.
The four-year project, funded through the U.S. Department of Agriculture’s Specialty Crop Research Initiative, will establish a Center of Excellence for Orchard Robotics in Cornell’s Department of Biological and Environmental Engineering. Researchers say the goal is to make orchard robots more capable, autonomous and affordable enough for commercial growers to use.
The work could have particular significance in New York, the nation’s second-largest apple-producing state and a major center for specialty crop agriculture.
Labor costs have become one of the largest pressures facing orchard operators. One large Washington grower participating in the project said labor represented about 45% of total costs 15 years ago and now exceeds 60%, while apple prices have remained largely stagnant.
Cornell researchers say automation could help growers reduce costs while also limiting worker exposure to injuries associated with ladder falls, repetitive motion and machinery.
Manoj Karkee, a professor in Cornell’s College of Agriculture and Life Sciences who is leading the project, said researchers want to automate as many repetitive orchard tasks as possible while creating new jobs involving the manufacturing, maintenance and supervision of robotic equipment.
The project includes researchers and industry partners from Cornell, Washington State University, Oregon State University, Penn State, Michigan State, Carnegie Mellon, Stanford University and Fresh Fruit Robotics, an Israel-based company.
Researchers are focusing on several challenges that have historically made orchard automation difficult.
Unlike crops such as corn and cotton, apples and other tree fruits vary widely in size, position and visibility. Robots need to distinguish fruit from leaves, branches, trellis wires and other obstacles while moving through orchards without damaging crops or equipment.
Recent advances in artificial intelligence and machine learning are helping researchers address those problems.
Karkee said neural-network-based systems are giving robots a much stronger ability to interpret orchard environments and make decisions. That includes determining where fruit is located, how it sits in relation to branches and other fruit, and where a robot can safely move.
Researchers are also continuing work on soft robotic hands capable of picking fruit without bruising it.
The project won’t focus solely on engineering. Economists, sociologists, horticulturalists and computer scientists are also involved because researchers say a technically successful robot will have little value if growers can’t afford it or don’t see enough benefit to adopt it.
Cornell Cooperative Extension fruit specialist Mario Miranda Sazo will work directly with New York apple growers to gather feedback and help ensure the technology addresses real-world needs.
Researchers will also develop digital versions of orchards for analysis, train artificial intelligence to evaluate tree canopies and determine which young fruit should be thinned, and study the economic and cultural factors that influence whether farmers adopt new technology.
One of the biggest questions is whether a single robotic platform can perform multiple jobs throughout the growing season.
Industry partners say a costly machine used only during harvest may not provide enough return on investment. A robot capable of pruning during winter, pollinating and thinning during spring, controlling weeds and harvesting fruit later in the year could be more practical for growers.
That multipurpose approach is expected to be a major focus of the Cornell-led project as researchers try to move orchard robotics from experimental prototypes into commercial operations.
Cornell researchers have already been testing prototype systems at Cornell Orchards, and project partners say rapidly improving AI technology could accelerate the timeline for broader adoption.
The grant-funded project is expected to run for four years, with researchers working toward systems that can operate more independently while meeting the cost and reliability demands of commercial fruit growers.
FingerLakes1.com is the region’s leading all-digital news publication. The company was founded in 1998 and has been keeping residents informed for more than two decades. Have a lead? Send it to news@fingerlakes1.com.