AI Comes to the Apple Orchard -- From Pollinating to Pruning to Picking: New robots promise to automate a process that has always relied on human workers -- and bees

"Apple orchards are largely untouched by automation. But that could be changing.

Recent breakthroughs in robotics and artificial intelligence could enable robots to do much of the work that orchards require, potentially transforming how apples get to a fruit bowl near you.

A number of universities and startups are developing specialized robots for each stage of apple production, from pollinating the trees to fertilizing them, pruning them and harvesting the fruit.

The push for automation comes as orchards face rising labor costs, and as they're evolving in a way that would mesh well with robot apple pickers. Modern growing techniques train trees into narrow, rectangular shapes of limited height -- rather than the natural big, bushy shape with a rounded canopy -- making it easier to reach the fruit.

Here's how robots are potentially revolutionizing the orchard, step by step.

Pollinating

Researchers have developed robotic pollinators to help where there are no local bees to do the job. Orchards where bees are lacking currently pay to have them brought in from elsewhere.

The robots use cameras to identify apple blossoms. When they spot a target, a mechanical arm moves in close and releases a precise burst of pollen through a nozzle. Early results are promising. In tests conducted by researchers at Washington State University, 84% of targeted flowers were pollinated.

Researchers envision using these systems to strategically pollinate flowers in the most accessible locations on trees at the optimal time. This could help improve robotic harvesting by ensuring fruit grows where machines can best reach it.

Pollinating only certain flowers this way would also limit the number of apples a tree produces, resulting in better-quality fruit. Orchard workers currently achieve this by thinning a tree's fruit manually.

Fertilizing

Apple orchards traditionally have applied uniform amounts of fertilizer to thousands of trees. New robotic systems developed at Washington State are changing this by treating each tree individually. They consider multiple indicators, including trunk diameter, canopy growth patterns and fall leaf-color changes, and use AI models to determine precise nitrogen requirements. A robot then delivers customized doses of liquid fertilizer to each tree's root zone.

Researchers believe precisely targeting individual trees could significantly reduce overall fertilizer use while maintaining consistent growth across varying soil conditions.

Pruning

A robotic pruning system has to overcome one of orchard automation's greatest challenges: Pruning requires complex choices about which branches to remove to optimize the next season's growth and fruit production.

Researchers at Washington State and Oregon State University are testing experimental pruning systems in commercial orchards. The robots must evaluate the entire structure of each tree, analyzing branch length, angles, thickness and spacing to determine which cuts will promote ideal growth patterns.

Harvesting

Harvesting apples is particularly challenging because the fruit bruises easily. Researchers have developed a couple of ways to pick apples without damaging them, using either suction or grippers.

Advanced.farm, a robotic startup based in Davis, California, has developed a robotic apple harvester now being tested in commercial orchards. The machine uses cameras to identify ripe fruit by color, then employs suction cups to pick the apples.

The harvester's arms can reach from two to 12 feet high, which is generally as high as orchard trees are grown. The machine typically collects about half an orchard's ripe apples, with human crews following up to pick fruit the robot missed, including apples hidden behind leaves or branches.

The system's AI analyzes data from each harvest to refine its movements so it can improve its picking accuracy while maintaining the gentle handling needed to prevent damage to the fruit.

Researchers at Washington State are testing a different approach to robotic harvesting, using inexpensive, fabric-based robots that have three silicone fingers that close around apples. The robots' softness means a bump from one of them won't harm workers or damage tree branches.

Several units work together from a mobile platform, with three or four robots positioned on each side. They target lower fruit, working alongside human pickers who handle higher clusters and more-complex picking situations like tangled fruit.

Currently being tested in lab conditions with fake trees and fruit, the system will face its first orchard trials during the next harvest season, in the fall. The developers hope the relatively low cost of these simple robots will encourage widespread adoption." [1]

1.  Artificial Intelligence (A Special Report) --- AI Comes to the Apple Orchard -- From Pollinating to Pruning to Picking: New robots promise to automate a process that has always relied on human workers -- and bees. Snow, Jackie.  Wall Street Journal, Eastern edition; New York, N.Y.. 03 Feb 2025: R6.