SurFloor
Future Project
Cherry Kim, Seojoon Hong
Yonsei University
Embedded Files
"SurFloor" aims to localize the visually impaired using floor surface characteristics. While floors of identical materials might seem uniform, we discovered that each location has unique levels and roughness. These differences are detected using an IMU sensor attached to a cane, enabling differentiation between locations.
Background
This project idea arose while walking through our department's corridor. We suddenly noticed that the floor is not completely flat, but rather has bumps and curves, as depicted in the left image.
Pre-processingÂ
We attached an IMU sensor to the white cane to capture floor bumpiness while dragging the cane. We followed the following steps to preprocess the data: First, we converted the digital sensor readings into physical acceleration values to better understand the data in real-world implications. Furthermore, considering that visually impaired individuals typically carry their cane diagonally, we calculated the vertical acceleration using the x-axis and z-axis values. These steps enable us to analyze and capture the physical displacement of the floor using an IMU sensor.
Preliminary Evaluation
For the preliminary evaluation, one individual slid his cane along each line five times. The results indicate that each line produces distinct signals, which are easily distinguishable from one another, while the signal remains consistent with repetition.
Future Considerations
What exact physical property causes different signals at each position?
How accurately can we estimate the position when applied in a real-world environment?
How can we adjust the data when the user sweeps the cane, which leads to noise in the IMU data?
Page updated
Google Sites
Report abuse