Microsoft creates smart, robotic mosquito traps

Microsoft is working on a new kind of mosquito trap that – when coupled with machine learning and gene-sequencing techniques – could help scientists identify outbreaks of diseases like Zika, dengue fever, malaria and West Nile virus before they happen.

Why it matters: Many mosquito-borne illnesses circulate in animals before making the jump to humans. If scientists can identify disease-carrying mosquitoes early on, they can take steps to control an outbreak and prevent it from happening.

How it works: The trap is a tower with 64 tiny doors, each capable of capturing one mosquito. Different attractants, like carbon dioxide and human body odors, can be used to lure different types of mosquitoes. The trapped insects' genes are then sequenced and analyzed by a machine learning algorithm to identify the species of mosquito, what animals they may have fed on and which diseases they carry.

The advance: Currently, scientists use sentinel animals to detect mosquito-borne diseases. They might place a coop of chickens in a known mosquito breeding area, and test the chickens regularly for West Nile Virus. But Microsoft's Project Premonition uses the mosquitoes themselves as sentinels, allowing them to can catch diseases that are spread through animals and insects other than birds, like lizards and horses. "Essentially, we're turning these mosquitoes into data-collecting devices," said Ethan Jackson, the project leader.

The bells and whistles: These traps don't just catch the mosquitoes: they also track the time of day they were captured and what the temperature and humidity were, data that could help determine the best time of day to spray insecticides. They even use a beam of light to measure the rate of the trapped mosquito's wingbeats – which lets them identify the species.