FingerIO: Using Active Sonar for Fine-Grained Finger Tracking
summary
Transforms a phone or smartwatch into an active sonar system that transmits inaudible 18–20 kHz OFDM pulses (48 kHz sampling, 5.92 ms spacing) and uses two microphones to track the closest moving echo. Demonstrates 8 mm median accuracy (0.8 cm, 1.2 cm max) on a Samsung Galaxy S4, a smartwatch prototype achieving 1.2 cm over a 25×50 cm area, works through fabric and in-pocket occlusion, uses the in-built speaker and microphones, and detects a near-phone start/stop swipe with >90% accuracy and only two false positives on the phone (zero on the watch).
AI-written summary · about
chapters (12)
- 0:00Introduction to FingerIO
- 1:54Limitations of radar-based tracking
- 2:59FingerIO approach and goals
- 4:03How sonar works on phones
- 5:51Challenges in sub-centimeter tracking
- 7:59OFDM for precise timing
- 10:31FingerIO system design
- 11:17Accuracy evaluation results
- 13:06Smartwatch prototype testing
- 14:07Gesture detection for start/stop
- 15:14Summary and future directions
- 15:36Q&A session
related talks
20mAccelerometers as an Instrument
16mGetting hands on with VR
19mHaptic Feedback to the Palm and Fingers for Improved Tactile Perception of Large Objects
22mIn two worlds at once: Social AR and VR
19mPulp Nonfiction: Low-Cost Touch Tracking for Paper
31mHybrid Body Craft
42mFeeling Films – The Future of Tactile Media
16mPentelligence: Combining Pen Tip Motion and Writing Sounds for Handwritten Digit Recognition