- Setup BBB
- Audio input
- Tensorflow lite
- Test keyword spotting examples
- Record Audio samples
- Record a few audio samples from 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315° at varying distances
- Combine STFT audio processing and fuzzy logic azimuth estimation
- STFT: See if we can identify frequecny spikes when speech is from a certain direction
- Fuzzy Logic: Once we identify some characteristics, we will use fuzzy logic to estimate direction of audio
- Combine keyword spotting and azimuth estimation
- Control Servo with azimuth results
- Create demo with realsense and base
This package utilizes a single microphone and a BeagleBone Black board to detect the relative location of human speech and automatically orient a Realsense camera towards the sound source. When a human voice is detected, the software calculates the direction of the sound and adjusts the camera's orientation using a servo.
- Digital Servo
- Common microphone
- Good microcontroller or SBC
- A USB camera
- WIP
- WIP
- WIP
See Work Referenced. Workflow improved by ChatGPT.
This project is released under the MIT License. See the LICENSE file for more information.