Echolocation – How drones can hear walls

Echolocation – How drones can hear walls


Can a drone hear the shape of a room? Or rephrased: Is it possible to precisely
measure the shape of a room using sound waves? The basic scenario is a room with flat walls,
and maybe a ceiling and a floor. Various microphones and a speaker are placed
in the room. Previous studies have already mathematically
proven that four microphones and a loudspeaker are sufficient to pinpoint the walls and also
calculate their inclination. Professor Gregor Kemper from the Technical
University of Munich and his colleague, Professor Mireille Boutin, who is working at Purdue
University in Indiana, therefore took the idea a step further. In their theoretical approach, they placed
a loudspeaker and four microphones on a drone. That makes it much more practical to basically
fly the equipment into the room. But it also made developing the experiment’s
mathematical framework much harder for the two researchers. The loudspeaker emits a high-frequency acoustic
pulse which the walls reflect. The microphones pick up the echoes. Next, the echoes must be assigned to the wall
from which they have emanated. In fact, this echo-matching is the main challenge. To accomplish this, Kemper and Boutin developed
a new algorithm. Once the echoes are sorted out, the positions
and the inclinations of the walls can be calculated using geometric methodologies akin to localization
using GPS signals. Unfortunately, these calculations can sometimes
go wrong. The relationships between echoes occasionally
fulfill mathematical constraints that seem to indicate they stem from a wall that isn’t
there, giving rise to so-called ghost walls. Now with four microphones put independently
throughout the room, this has been shown to be almost impossible. However, when the microphones are mounted
in rigid positions on a drone, it should be expected that ghost walls will occur with
greater probability. That’s because compared with the setup of
microphones distributed freely in a room, microphones mounted on a drone are limited
in their freedom of movement. But by setting up a more elaborate mathematical
framework, Kemper and Boutin were able to demonstrate that a drone will assume with
near certainty a position that avoids ghost walls. The researchers hope their scenario will become
even more realistic in the next step: They are developing a mathematical solution to
handle errors and inaccuracies in measurement. They also intend to address the theoretical
placement of the loudspeaker and microphones on ground-based vehicles.

Leave a Reply

Your email address will not be published. Required fields are marked *