Pollination Robots: A New Frontier in Agriculture Robotics

Robotics has changed the facet of farming in recent years with its revolutionary capabilities. From strawberry picking robots to automatic weed plucking, robots have helped fill the labor shortage in the agriculture industry. As robots are becoming a key part of modern automated farming practices, scientists have found yet another application of robots in this domain — Pollination robots! As the name suggests, these droids will enable the pollination of plant flowers both in farms and in the wild.

All flowering plants need pollination for reproduction. In planets, the male flower parts, called stamens, generate pollen, which gets fertilized by female flower parts, called pistils, to form seeds. The stamen discharges pollen directly onto the pistil in self-pollinating flowers. However, in cross-pollination, pollen movement occurs from one plant to another, which depends on vectors like bees, butterflies, dragonflies, etc. When bees and other insects eat on flowers, pollen sticks to their bodies and is then deposited on the next plant they visit. Cross-pollination is an important part of planet reproduction as they help in enhancing genetic diversity. It also helps in improving the quantity and quality of crops.

Bees are responsible for around 80% of all pollination on the planet. Bees pollinate more than 70% of the top 100 crops, which provide roughly 90% of the world’s nourishment. Sadly, today their lives are in danger. Researchers have noticed that the population of bees is declining due to Colony Collapse Disorder, Varroa mites, use of neonicotinoid pesticides, and loss of habitat. One chemical compound found in pesticides, called Thiamethoxam, is said to prevent queen bumblebees from laying eggs. According to predictions based on a mathematical model, this could lead to the extinction of local wild bee populations. In 2016 alone, the United States reported a loss of 44% of all honeybee colonies. Moreover, seven species of yellow-faced bees in Hawaii are on the verge of extinction. 

To solve the impending danger of acute food shortage due to the dwindling number of bees, scientists have turned to robotics to help address such woes and minimize dependency on bees. Thanks to advancements in the deep learning algorithm of artificial intelligence, robots can quickly and accurately identify flowers for pollination. 

Read More: Artificial Intelligence Powered Precision Agriculture Can Revolutionize The Food Production Industry

Arugga AI Farming, an Israel-based company, develops robots that can automatically go down a greenhouse row, leveraging artificial intelligence and cameras to recognize tomato flowers that are ready to be pollinated. The robot features an air nozzle system that sprays a calibrated pulse of air on the selected flowers, ‘shaking’ them to mimic bumblebee buzz pollination. Unlike some other crops, Tomato blossoms pollinate themselves when vibrations shake pollen loose, eliminating the need for cross-pollen transmission between flowers. 

Apart from pollination, these pollination robots reduce the risk of disease transmission because the air blast pollinates the flowers without coming into contact with them. Other advantages this technology has over the bumblebee include its capacity to work efficiently under artificial lighting and in extreme hot or cold conditions, both of which are detrimental to the insects’ pollination performance.

This year, Arugga has raised US$4 million in pre-Series A funding. The round was led by Cresson Management, with Biobest Group, Terra Venture Partners, Equicelar, and Smart Agro, among the other investors, to take part. 

In 2018, inspired by bees, a team of West Virginia University researchers built BrambleBee, an autonomous robot that can pollinate bramble, blackberry, and raspberry plants, in a greenhouse setting. This robot uses cutting-edge LIDAR mapping and localization techniques, along with additional visual perception, path planning, motion control, and manipulation tools. LIDAR operates similarly to radar, except instead of sending out radio waves, it produces pulses of infrared light (called invisible lasers) and detects how long it takes for them to return after colliding with adjacent objects.

BrambleBee first obtains up-to-date information on flower cluster locations and pollination readiness by doing an ‘inspection pass’ of the greenhouse. During this, its onboard camera detects adjacent flower clusters. The identified clusters’ positions are subsequently documented on a map of the plant rows. Bramblebee then selects where it will pollinate flowers once this initial assessment step is over. This choice is made by balancing the number of accessible flower clusters that are ready for pollination while reducing the distance driven. 

When BrambleBee deems that a flower is ripe for pollination, it will gently stroke the blossom with a tiny 3-D printed brush containing flexible polyurethane bristles on the end of its arm. In this way, pollens are transferred from the anthers to the pistils. BrambleBee robot uses precise motions to maneuver this mechanism, spreading pollen into pistils while avoiding damage to the flowers. The robot even remembers which flowers it has already struck, allowing it to proceed with other plants in the greenhouse.

The above-mentioned examples are not the only way of pollinating crop flowers. For example, pollination might be outsourced to automated drones that carry pollen grains to specific flowers, according to materials scientist Eijiro Miyako of the Japan Advanced Institute of Science and Technology in Nomi. While his initial plan entailed rubbing grains into flowers with a pollen-coated drone, he quickly discovered that the procedure was damaging the blossoms. 

Last year, while playing bubbles with his son, Miyako realized a solution to perfect the drone problem. He concluded that his team could use pollen-laden soap bubbles sent from the drone and gently land on the blooms for a more gentle approach. The researchers used this approach to hand-pollinate pear trees in an orchard to evaluate the viability of their pollen-loaded bubbles. According to the researchers, the trees produced almost the same amount of fruit as plants pollinated by manual pollination.

A start-up from the United States called Dropcopter uses unique pollination delivery technologies and specialized drones with cutting-edge algorithms. This drone is already pollinating apples, almonds, cherries, and pears at a rate of 20 acres per hour per drone. The average pollination cost per acre across the board, according to the company, is around US$300. 

Bees are vital to ecosystems and agricultural pollination, which is why the IUCN (International Union for Conservation of Nature) is urging immediate investment in new research to find methods to reverse this decline. Pollination robots may not replace bees nor restore ecological balance but can certainly address the pollination crisis. 

While pollination robots may seem like an avant-garde solution, it is still not economically feasible on a large scale or at least 4-5 years away from commercial release. Furthermore, even the existing models need more fine-tuning to be able to pollinate without harming flowers, enable more cross-pollination, etc. However, the silver lining is that these pollination robots offer something that bees don’t — data. The data gathered while attempting to pollinate plants can be utilized to boost agricultural practices.