Feng Zhang

LiDAR-Based Gesture Recognition and Multi-Sensor Fusion for Echo Devices

• Developed and validated a deep neural network to recognize human gestures using cost-effective LiDAR sensors.
• Engineered a multimodal fusion framework that combines ultrasound and WiFi CSI data to enhance presence detection.

Integrated and Evaluated mmWave Radar Signals for Autonomous Systems

• Led improvements in radar data filtering, representation, modeling, and performance bounding.
• Integrated radar data into the perception DNN, boosting system accuracy in challenging scenarios.

Developed Sensing Features for Echo Family Devices

• Built machine learning algorithms to recognize human activity patterns using WiFi CSI, mmWave Radar, and thermal sensors.
• Developed comprehensive test plans to validate performance and ensure product robustness.

Invented a Series of WiFi Sensing Systems

• Created a robust, calibration-free WiFi motion detector with near-zero false alarms and through-the-wall coverage.
• Work highlighted by Engadget; commercialized on Linksys routers; contributed to CES 2020 Best Innovation Award.
• Developed the first practical WiFi sleep monitoring system using machine learning.
• Designed a super-resolution WiFi imaging system reusing a single 60GHz WiFi radio as a millimeter wave camera.