A new open-source AI tool is being used to power a laser turret to go after some of humanity’s most annoying pests, cockroaches. In a study published in Oriental Insects a week ago, Ildar Rakhmatulin, a researcher at Heriot-Watt University, and fellow researchers combined machine learning with machine vision in a series of experiments on domiciliary cockroaches. In the experiment, not only was the AI able to detect the roaches, but they were able to zap them at an impressive distance of 1.2 meters or almost four feet away.
This isn’t Rakhmatulin’s first journey combating insects with the power of technology. In some earlier projects, he used a Raspberry Pi that attached to lasers to go after summer’s greatest villain, mosquitoes, according to Vice. But in this latest project, he used a more powerful computer system that allowed for greater precision detection. Rakhmatulin explains, “I started using a Jetson Nano that allowed me to use deep learning technologies with higher accuracy to detect an object.”
But how did he and the team accomplish what could be the cockroach’s biggest threat? Well, as Rakhmatulin explains, it was thanks to the Jetson Nano. The device is a small computer that runs machine learning algorithms. Empowered by these algorithms, the computer is able to process digital signals from two attached cameras to detect and locate the cockroach. Then, it transmits the information to a device called a galvanometer. This device, which was used to find faults in telecommunications cables, measures electric currents and tells the laser where to shoot.
This experiment didn’t just have the team firing away and blasting insects. According to the same paper, they were also able to configure the device at differing power levels. Because of this, they found that they could influence the behavior of the insects by what was essentially scaring them by triggering a flight response. At the higher levels, the cockroaches were fried.
For those who are sick of dealing with roaches and may not want to bother with poisons and other chemicals, Rakhmatulin has good news for you. “I use very cheap hardware and cheap technology and it’s open source…All sources are uploaded in my GitHub and see how to do it and use it.”
Because this is open source, others have already experimented on other pests and Rakhmatulin sees the technology’s benefits. “If it can damage cockroaches, it can also damage other pests in agriculture.” This is a quite interesting use of machine learning and one that could see laser turrets in homes. Though, it should be a sight if that happens.