In a world where virtual meetings are the norm, muting microphones to avoid interruptions is simple. But what about real-life gatherings? Researchers at the University of Washington have introduced a groundbreaking solution: a shape-changing smart speaker with self-deploying microphones.
Locating and Controlling Sound in Real-Life Environments
The challenge of isolating and controlling sound in crowded spaces without visual cues has long perplexed researchers. This new system employs self-deploying microphones and deep-learning algorithms to divide rooms into speech zones and track individual speakers’ positions.
The system allows users to mute specific areas or separate simultaneous conversations, even if the speakers have similar voices. These microphones, resembling small Roombas, automatically deploy and return to charging stations, offering mobility and ease of setup in various environments.
Image credit: Itani, et al, Nature Communications
The Breakthrough Published in Nature Communications
The University of Washington team, whose research was published on September 21 in Nature Communications, overcame previous limitations. Past research on robot swarms required overhead or on-device cameras, projectors, or special surfaces. This system, however, is the first to distribute a robot swarm accurately using sound alone.
How It Works
The prototype consists of seven small robots that navigate tables of various sizes using high-frequency sounds and sensors to avoid obstacles. These robots deploy themselves for maximum accuracy, allowing for better sound control than manual placement. They position themselves at a distance from one another, making it easier to differentiate and locate speakers in a room.
Impressive Test Results
Testing in offices, living rooms, and kitchens with groups of three to five people speaking, the system accurately discerned different voices within 1.6 feet (50 centimeters) of each other 90% of the time. It processed three seconds of audio in an average of 1.82 seconds, making it suitable for live streaming, although a bit slow for real-time communications.
Future Possibilities
Researchers envision deploying acoustic swarms in smart homes to enhance interactions with smart speakers. For example, only individuals within a designated “active zone” on a couch might control a TV with their voices.
Privacy and Safeguards
The technology, while exciting, raises privacy concerns. To address potential misuse, the microphones rely on sound, not onboard cameras, are visible and feature active indicators when operational, process audio locally for privacy, and can be used to create private zones.
This innovative system represents a significant leap forward in sound control technology, with applications spanning from personal privacy to enhanced smart home interactions. As the technology evolves, it may bring us closer to science fiction concepts like the “cone of silence.”
Publication:
-
Itani, M., Chen, T., Yoshioka, T. et al. Creating speech zones with self-distributing acoustic swarms. Nat Commun 14, 5684 (2023). https://doi.org/10.1038/s41467-023-40869-8
Source: UW
