
Track wild animals with smart microphones
This continuous data stream is very helpful to detect trends in an ecosystem more precisely than just taking measurements once per year
Any nature lover will tell you: the best way to identify the species present in a natural habitat is to listen attentively to their symphony of roars, growls, squeals and squawks. Synature, an EPFL startup, has created a smart microphone system that can record these sounds automatically over a vast area and analyze them using artificial intelligence. Its technology gives organisations, nature stewards and consultants a powerful new tool for assessing the impact of changes to natural habitats – such as the creation of a natural reserve or the construction of new buildings – expanding their research on the environment and animal behavior, and helping public officials and economic policymakers make informed decisions. “Our devices also provide an acoustically engaging way to get people interested in nature,” says Synature CEO Olivier Staehli. The system has been tested in Glaris canton, France and South Africa , and will soon be ready for larger-scale manufacturing.
Three-month battery life thanks to a reduced power requirement
Using animal sounds to monitor wildlife populations is nothing new, but it’s done manually: researchers use microphones to collect recordings on site, and then analyze the data later. “Our smart microphones filter out the background noise from the animal sounds,” says Staehli. “Then algorithms recognize the sounds made by specific species. This information is displayed in an app in an easy-to-read format.”
Unlike other wildlife monitoring methods such as remote sensing and eDNA – DNA present in the environment, derived from living organisms via their cells, excretions, hairs, scales or organic debris-, Synature’s system records continually. This continuous data stream is very helpful to detect trends in an ecosystem more precisely than just taking measurements once per year, for example. As a result, the challenge for its developers was to find a way to cut the amount of power used by the microphone so as to maximize its battery life. This is crucial to reduce the required field work to a minimum. To detect animals they use a novel transformer-based architecture that enables more accurate detection with less data than more traditional methods. Synature is expanding its system to detect amphibians, alongside thousands of bird species.
