MeMic, The AI Wearable That Only Listens and Records Its Owner Speak Powered by XIAO nRF52840 Sense
I have been a huge fan for AI-powered wearables that record audio continuously, as they have various applications such as health monitoring and cognitive augmentation. However, privacy concerns surrounding always-listening devices have hindered the adoption of many AI-powered tools.
What if I tell you that there is a wearable device that activates its microphone only when its wearer speaks? To addresse the privacy issue, team of engineers consisting Cayden Pierce, Wazeer Deen Zulfikar, and Pattie Maes at MIT Media Lab have developed MeMic, a wearable device that actually activates its microphone only when the wearer speaks. This innovative solution aims to make wearable recorders more socially acceptable by easing privacy worries.
MeMic can be incorporated into a variety of wearable devices.
Credit: W. Zulfikar et al.
Hardwares
To build the MeMic device, the following hardware components are used:
-
- Seeed Studio XIAO nRF52840 Sense (with 6-axis IMU and RGB LED onboard)
- Small LiPo battery
Components of MeMic
Credit: W. Zulfikar et al.
How the Project Works
MeMic operates by leveraging XIAO nRF52840 Sense’s onboard 6-axis Inertial Measurement Unit (IMU) that detects vibrations from the wearer’s body. When the wearer speaks, the IMU collects motion data, which is processed by a locally run algorithm on the XIAO nRF52840 Sense microcontroller. This algorithm filters signals to voice-related frequencies and analyzes z-axis acceleration data to determine if the wearer is speaking. Upon detecting speech, MeMic activates the microphone and turns on the RGB LED light to indicate active recording, ensuring transparency. The device has been tested in various wearable forms, such as pendant necklaces and smart glasses, with positive feedback on its privacy-friendly design.
“We needed to build a wearable with audio, IMU, power management, and Bluetooth in a tiny lightweight form factor, and we wanted to prototype it fast. Using multiple different off the shelf modules connected together would have been too bulky, and a custom PCB was overkill and too slow to iterate. The Seeed XIAO nRF52840 Sense board had everything we needed already integrated into a tiny lightweight form factor, so it was the obvious choice.”
Call To Action
While MeMic may alleviate some concerns, it is not without its flaws. For instance, it will still pick up the speech of bystanders if they talk simultaneously with the wearer. Additionally, MeMic requires a tight mechanical coupling between the device and the user, which may become uncomfortable during extended use. Moving forward, the engineer researchers plan to explore contact-based microphones and conduct longer-term trials to address these issues. For detailed info check out their paper on ACM Digital Paper or read the detailed project feature by Hackster.io.
End Note
Hey community, we’re curating a monthly newsletter centering around the beloved Seeed Studio XIAO. If you want to stay up-to-date with:
🤖️ Cool Projects from the Community to get inspiration and tutorials
📰 Product Updates: firmware update, new product spoiler
📖 Wiki Updates: new wikis + wiki contribution
📣 News: events, contests, and other community stuff
Please click the image below👇 to subscribe now!
About Author
