Self-touch gesture recognition via a novel magnetic sensing platform — nuanced facial touches, subtle finger scratches, and beyond.

First attack scheme that simultaneously listens and injects on commodity keyboards via magnetic sensing.

Overcomes initial parameter sensitivity of the LM algorithm, expanding magnetic sensing usability for mobile scenarios.

Backward compatibility of the BLE v5.1 direction finding feature with legacy BLE devices.

Real-time 3D tracking of passive magnets embedded in tabletop objects enabling rich interaction.

World's first invisible fiducial tag for robots using magnetic sensing — enabling pose calibration and navigation without cameras.

First wearable system for drift-free joint tracking. Accuracy: 2.72°, 4.14°, 4.61° for elbow, knee, and shoulder.

Deployment of magnetic tracking on commodity Android smartphones — energy-efficient and highly effective.

Magnetic encoding of road markings that remain visible regardless of weather — tested on real-world roads.

Leveraging widely available smartphone sensor data to infer vehicle type — a new privacy leak in driving apps.

MAGIC: systematic framework to automatically calibrate soft- and hard-iron disturbances for a MEMS magnetometer array.

First software-defined PHY enabling agile reprogrammability in wireless backscatter networks.

First untethered, mobile magnetic tracking platform for fine-grained hand tracking — AR/VR interaction and health monitoring.

B-Auth: driver authentication via battery voltage patterns correlated with driving behavior.

Open-source vehicle-agnostic end-to-end framework for vehicular data collection, translation and sharing.

IoT + AI framework for detecting safe left turns — democratizing traffic safety information for navigation apps.

Automated translation of in-vehicle CAN data — making the data flow in cars transparent to developers.

LocBLE: fine-grained location of BLE beacons using commodity smartphones. Collaborative work with Hewlett Packard Labs.

V-Sense: vehicle steering detection middleware running on commodity smartphones with no additional sensors.

Sniff and power attacks exploiting plaintext CSI feedback in MU-MIMO — and defenses against them.