CoNEXT '18 Artifacts
Paper #32: Boosting fine-grained activity sensing by embracing wireless multipath effects
This repository contains algorithms for enhance fine-grained activity sensing using “virtual” multipath. The fine-grained activities include human respiration monitoring, finger gesture recognition, chin movement tracking when speaking. We employ WARP platform to collect data. The original signal does not show obvious variations while the target performs each activity. After adding extra static multipath, we can change the sensing signal with good sensing capability.
The data locate in the directory /data/.
- respiration.mat is the human respiration Wi-Fi signal for 1 minute.
- finger1.mat is the Wi-Fi signal for finger gesture 'up'
- finger2.mat is the Wi-Fi signal for finger gesture 'no'
- chin.mat is the Wi-Fi signal for chin movement when speaking 'how are you, I am fine'. chin.m4a is the audio groundtruth, which can be loaded automatically.
To run the code, you only need to change the variable fname in Line 6 to the corresponding dataset name, then run the main function.
If you have any questions, please contact [email protected]