A group of engineers have now created what they call an acoustic swarm. This swarm of intriguing little robots – which move around like little Roombas – can wheel around an area, helping to control the separation of audio from multiple conversations in one room. Essentially, these swarms could let you mute other people’s conversations and focus on your own.
Now, we’ve seen similar tech – at least as far as muting tech – appearing in headphones and other smart devices, like the iPhone’s voice isolation. This tech relies on filtering out background noise for one person. This acoustic swarm, however, focuses on creating a cone of audio around different groups, allowing it to control where audio from multiple groups and conversations is going.
It is obviously a bit more involved than the first example I used, which relies on background noise filtration. And, if the video that the engineers have used to showcase the technology is any indication, it could really help out at big gatherings of people, at least once the tech is better updated and improved upon.
The acoustic swarm (that really is a cool name for a bunch of audio-based robots) is self-deploying. All the speaker has to do is arrange the seven microphones that the robots use to create a partition within the room, acting as a “speech zone.” The robots then use this zone to track and identify the voices, even as they move.
One of the biggest purposes behind the design is to allow you to mute noisy zones – like annoying people in the background – when you need to for important conversations, video conferences, and other things. It’s a really nifty idea that you can read more in-depth about in a paper on the design featured in Nature Communications.
The engineers say that they have tested the acoustic swarm in multiple real-world experiments, including within kitchens, offices, and other locations where between three and five people were talking, with different conversations going on. The researchers also say that the robots had no prior exposure or knowledge of the locations of the voices they would be responding to.
Ultimately, the research says that the robots could localize the voices 90 percent of the time, within 1.6 feet of each other. That, I have to admit, is very impressive, and I’m absolutely intrigued to see where it goes from here.