This Raspberry Pi AI clock listens and smells for the current time
Raspberry Pi clocks have been a staple in the maker community, but the latest projects show an increasing level of complexity. Max Björverud’s latest creation is a prime example: Raspberry Pi-powered clocks that use AI to estimate the current time based on smell and sound.
Max Björverud has developed two unique clocks, each utilizing a different input method:
- The Ear Clock: Uses a microphone to detect audio cues.
- The Nose Clock: Uses a multichannel gas sensor to detect smells.
Both clocks function independently, gathering data through their respective sensors and using AI to estimate the current time. Over time, they improve their accuracy based on accumulated data. Despite this, the primary aim is to highlight the inherent uncertainty in AI-driven predictions.
Technology and Tools Used
- Raspberry Pi Models:
- Ear Clock: Powered by a Raspberry Pi 3B+.
- Nose Clock: Powered by a Raspberry Pi 4 B.
- Sensors:
- Microphone: For audio detection in the Ear Clock.
- Multichannel Gas Sensor: For smell detection in the Nose Clock.
- Software:
- AI and Machine Learning: Utilizes Scikit-learn, an open-source Python library, which provides a variety of tools for machine learning.
- Visual Output: Uses openFrameworks, an open-source C++ toolkit for creative coding, to control the matrix panel displays.
How It Works
- Data Collection: Each clock collects data via its sensors. The Ear Clock listens to ambient sounds, while the Nose Clock samples the air for specific gases.
- Time Estimation: The AI processes this data in real-time to estimate the current time. The clocks improve their estimations over time by learning from the data they collect.
- Visual Display: The estimated time is displayed on matrix panels, with rows indicating the hour, minute, and second. Additionally, the bottom panel shows shapes based on recent data metrics.
While the accuracy of these clocks may be questionable, Björverud emphasizes that this uncertainty is intentional. The project aims to raise awareness about the nature of artificial intelligence and its reliance on data interpretation. Regardless of accuracy, the clocks provide a visual and conceptual representation of how AI can interpret environmental cues.
Max Björverud’s AI-driven Raspberry Pi clocks are a fascinating blend of art, technology, and commentary on artificial intelligence. By using sound and smell to estimate time, the project challenges conventional notions of timekeeping and underscores the limitations and potentials of AI in interpreting our world.
Read more: This Raspberry Pi AI clock listens and smells for the current time