Searching for survivors at disaster sites is one of the most common uses for drones. However, if these people were buried under the rubble, they would not be visible from the air. That's where LUCY comes in, allowing drones to find survivors by sound. Developed by scientists at the Fraunhofer Institute for Communication, Information Processing and Ergonomics in Germany, LUCY is an inexpensive MEMS (micro-electromechanical systems) microphone array that can be installed on the chassis of existing multi-rotor drones. Its name is an acronym for "Listening System Using Crow's Nest Array."
"The highest lookout point on a ship that can be seen in all directions is called the 'crow's nest,'" said Macarena Varela, who led the research with Dr. Marc Oispuu. "The same is true for LUCY - our system has almost no limitations and can hear sounds from all directions."
In the current version of the system, 48 powerful microphones are configured in a special geometry that optimizes the device's ability to pinpoint the direction of a specific sound source. The microphones can even capture frequencies that are inaudible to the human ear.
In other words, the artificial intelligence algorithm running on the LUCY signal processing device can block interfering sounds such as wind sound, rescue equipment, and the drone's own rotor. At the same time, the algorithms also search for sounds such as shouts, banging or clapping, which could be used by trapped survivors to attract attention.
Once these sounds are detected and their geographical origin is determined, the system can forward the coordinates to those on the ground who will perform the actual rescue. In fact, LUCY can also be used on the ground, without the need for drone loading.
Varela, Oispuu and their colleagues are currently developing an improved version of the system that will be more powerful and have 256 microphones.
Source: Fraunhofer
Compiled from:ScitechDaily