UOW AI mosquito detection device

Researchers at the University of Wollongong, led by Associate Professor Kiran Trivedi, unveiled a low‑power TinyML device that identifies malaria, dengue, Zika and other disease‑carrying mosquitoes by analysing the pitch of their wing‑beat sounds. The prototype, built from 40 GB of audio recordings and compressed to run on a tiny chip, was showcased at the UN AI for Good summit as a potential tool for public‑health surveillance in endemic regions.

A few days later the team reported test results, noting the model achieved 88.3 % accuracy on publicly available recordings. They highlighted ongoing work to improve noise filtering so the system can function reliably in noisy field conditions. The device operates offline and is positioned as a low‑cost alternative to traditional laboratory‑based mosquito monitoring, aiming to support remote and developing communities in tracking vector‑borne diseases.