In a sunny farm in AlUla, an ancient oasis city in northwestern Saudi Arabia, a killer AI-powered nano-sensor is on the prowl. Once its AI engine has locked on to its target, it can help detect a ravenous-type pest – the red palm weevil. 

A crackle, a puff of smoke, and the target are dead? Not really. 

Here the machine learning algorithm captures weevil larvae-specific sounds – distinguishes the munching noises from other noises – and detects infestations before they become visible.

In the last few years, date palm farms have been destroyed by the red palm weevil, a tiny insect, causing millions of dollars in losses in the region annually.  

The pests spend 80% of their lives inside trees and are almost invisible until it’s too late, leaving farmers with few defenses aside from drenching their fields in pesticides or felling and burning the infected trees. Of late, the red palm weevil is becoming resistant to pesticides.

Now, researchers from King Abdullah University of Science and Technology (KAUST) – Dr. Islam Ashry, Dr. Chun Hong Kang, and Prof. Boon S. Ooi – have developed an AI-powered sensing device to monitor and prevent weevil infestations. KAUST has launched a startup, called AK-Sens, to commercialize it, and currently in large-scale pilot in AlUla.  

HOW THE TECH WORKS

“Our approach involves sending laser pulses from a sensing device into an optical fiber, which can be wrapped around the trunks of multiple trees over a vast area. Sound interacts with the light signal inside the fiber, changing its frequency,” says Ashry.

“The fiber feeds the data back into the sensor that, with the aid of machine learning, separates weevil larvae sounds from background noises such as wind and birds and then indicates which trees are healthy and which are infested. In other words, we have developed a machine learning algorithm that can be taught to distinguish the munching noises from other sounds,” adds Ashry.

In the Middle East, agriculture represents 13% of the region’s GDP, and out of the region’s total population of over 400 million people, roughly 100 million, or 25%, are entirely dependent on agriculture, which is particularly vulnerable to pests. 

It took about four years for the team to build their technology. They started experiments in a lab, followed by outdoor farms in Al-Ahsa during a collaboration with the Center for Palms and Dates in Al-Ahsa and the Ministry of Environment, Water and Agriculture (MEWA).  

After successful trials in Al-Ahsa, they installed the sensor at the TADCO farm in Tabuk – a project funded by NEOM and AlUla.

“Data gathered by the installed sensors are helping teach the AI algorithm to filter out noise better,” says Kang. 

Soon, they will install their system in two more locations in the kingdom. The sensor can be in a roaming vehicle to pick up signals from the optical fibers surrounding thousands of trees.

PERVASIVE INFESTATION

The weevil infestation is pervasive in the region. It first appeared in Saudi Arabia and the UAE in the mid-1980s. Today, date palm farmers in the southwest of Saudi Arabia, Jordan, Oman, Israel, and the Palestinian Authority Territories are increasingly at war with tiny, unfamiliar enemies.

Although the red palm weevil comes from Southeast Asia, the insects can fly up to 50 kilometers daily.

“A single bug lays an average of 200 eggs that hatch inside the palm’s trunk within three days. The hatched larvae munch on the tree’s insides, and if not caught in their very early stages, the bugs kill the tree and spread to its neighbors,” says Kang.

“Date palm trees face certain death unless the infestation is discovered early in the insect’s life cycle. In Gulf countries and the Middle East, $8 million are spent yearly to remove infested palm trees,” he adds.

PREVENT THE USE OF CHEMICALS

As the weevils infest trees around the region, they release carbon dioxide into the air and accelerate global warming. Typically, farmers treat date palms with preventive pesticides a couple of times a year, which kills valuable pollinators like bees and releases chemicals into the surrounding environment. But AI monitoring can reduce the farmers’ use of these preventive chemicals by half. 

“Depending on the trunk’s circumference and distance between palms to palms on a farm, we will scan about 400 to 600 trees in one trial. Regular scanning could curb the pest in time and save palm trees from dying,” says Ashry.

Ooi says early detection of the presence of larvae is challenging because date palms do not show signs of distress until it is too late to save them. “We provide an early detection alarm system to alert farmers of infested trees when the larvae are less than two weeks old.”

The red palm weevils threaten thousands of hectares of land. And the risk to Saudi’s economy could reach nearly half a billion dollars if the problem is left untreated. “The total production of dates in Saudi Arabia is 1.55 million tonnes yearly, representing 17.3% of the total production of dates worldwide. The kingdom exports dates to more than 113 countries worldwide, bringing over $300 million,” says Ooi.

Red palm weevil infestation is a global problem, and it will threaten the livelihood of more than 50 million farmers. “We want to extend our tech in the kingdom and branch out to other countries in the region and eventually the entire world,” says Ashry.

The technology could also be used for earthquake detection, and infrastructure and wind turbine monitoring.

“The technology is scalable and can save lots on costs in the long run. It is also much more sensitive than the detection approaches currently used on farms,” adds Ooi.