During this week, our team was working on a completely different project- a kinetic sculpture that interacted with a user's movement. The project was very mechanically focused and was based off this kinetic sculpture. We quickly realized that we were interested in a more computational/electrical project, and so we pivoted to make the ASL interpreting gloves.
During week 2, we explored different methods of sensing hand movement and read a variety of papers [1, 2, 3, 4] about similar glove systems to learn about implementation. We purchased (and burnt out) our first accelerometer and started working on finding a glove that would be suitable for multiple users. Eventually, we found a simple, knit black glove that could stretch to fit different sized hands.
During week 3, we purchased flex sensors to detect the bend in each finger on the hand, and sucessfully read three different hand positions using basic Arduino code. At this point, the flex sensors were taped to the glove on the outside of the fingers (top of hand), and the glove was pretty comfortable to wear. This is shown in the first video on the right.
During the fourth week, we focused on making the glove more comfortable and usable. We designed rings, caps, and a bracelet to house the flex sensors and accelerometer (see the mechanical documentation). However, we eventually decided that sewing the flex sensors inside the glove on the bottom of the hand was the most comfortable. Small pieces of fabric created runners for the flex sensors to move and prevented the sensors from touching the skin.
During week five, we worked heavily on the computational aspects of the glove. First, we restructred the code so that it could more efficiently handle the large array generated by 26 letters. Additionally, we wrote a script that allowed users to calibrate the glove to their specific way of signing letters. Finally, we worked on printing the signed letters to an LCD screen so that people could use it without a computer. This is shown in the second and third videos on the right.
During week six, we worked on integrating the accelerometer with the rest of the code and re-designed the bracelet holding the electrical components to fit newer components. Additionally, we worked on integrating a Bluetooth receiver into the design so that users could use an Android phone rather than and LCD screen.
During week 7, we worked on finalizing the accelerometer integration with a new accelerometer and improved the integration with the Bluetooth receiver. We also worked on saving the calibration data to the Arduino so that one person could use it multiple times without having to re-calibrate it, however we were unsucessful due to the size of the array we were looking to store on the Arduino.
During the final week, we worked on making the sensor readings more accurate, implementing a duplicate letter function, a space "letter", and a new line function. We also worked on making the whole system battery powered for easier use. The final video on the right shows our final product.