This week, I completed the following items:
1. Researched, finalized, and ordered materials for the neck angle-calculation module. These included the gyroscope/accelerometer module, an ESP32-C6, and a battery to power these components. I also researched the communication protocols that I will need to configure and use to send data from the sensor to the ESP32 (I2C), and from the ESP32 to the main microcontroller (BLE).
2. Researched objective measures for determining poor neck posture. I hope to implement an approximation of craniovertebral angle, which is actually used for diagnosing forward head posture, but this will require good positional calculations from the accelerometer, which may not be feasible. To achieve this, I researched methods of improving calculation accuracy for angle/position, including Kalman filtering. An alternative threshold for neck strain is simply a 15 degree downward rotation of the neck.
3. Updated the design of the neck module to use a cap/visor instead of a neckband to improve wearability, comfort, and ease of component layout.
4. Tested and did initial calibrations on the OpenCV Python script that we will be adopting for the blink detection part of this project. From this testing, I confirmed that rudimentary blink detection worked, and identified areas that we will need to improve upon and modify to meet our use-case requirements. These included support for single-eye blink detection, ensuring that our rate-calculations do not factor in time where neither eye is in view of the camera, improving functionality in dimmer lighting, and improving landmark detection for glasses-wearers. However, as difficulties with running a Python script from the browser extension arose, Cora will be working on converting the script to JavaScript using OpenCV.js.
5. Prepared slides and presentation for the upcoming design review.
A couple of my deliverables from last week were not completed unfortunately (figuring out the Bluetooth implementation between ESP32/RPI, and drafting code for gyroscope calculations), as I spent much more time than expected on researching and selecting a microcontroller, and working on the design review presentation. However, these two items would be more effectively completed with the parts in hand, so this isn’t super detrimental to our schedule.
Deliverables for next week:
- Present design review and write design report.
- Set up the development environment for flashing the ESP32.
- Research how to implement bluetooth communication between ESP32 and RPi. If ESP32 arrives on Mon/Wed, work on actually setting up this connection too.
- If sensors arrive as well, start setting up the I2C interfaces and test sensor reliability.