Gesture Recognition Using Ambient Light :
- Built a novel machine-learning-based gesture recognition system only using a user’s shadow under indoor lighting.
- Achieved recognition accuracy of 96.36% across 15,000 samples from 20 users for 5 gestures.
- Published at ACM IMWUT 2018 Vol 2 Issue 1
Accurately decoding MIMO Streams for Visible Light Communication:
- Designed a novel data-driven (vs model-based) approach for decoding MIMO streams of indoor visible light communication.
- Demonstrated an order of magnitude lower Bit Error Rate than conventional channel matrix approaches.
Multi-User Activity Recognition Using WiFi
- Devised a multi-user activity recognition system using patterns from WiFi signals.
- Achieved recognition rate of 95.0, 94.6, 93.6, 92.6, and 90.9% for 2, 3, 4, 5, and 6 simultaneously performed activities from upto 6 users.
- Published at ACM MobiSys 2018, Munich, Germany