Recognizing Keystrokes Using Wi-Fi Signals

Executive Summary

When a user moves their fingers to type, it creates a disturbance in the Wi-Fi signal that is unique and can be determined by analyzing the CSI-Waveform from the Wi-fi card.  Our technology is able to detect and identify small movements of the hand by using a commercial Wi-Fi router and laptop.  This technology can be used to determine what is being typed on a computer, user identification by typing patterns, and many other applications of movement detection beyond finger movements. 

Description of Technology

The unique disturbances of the WiFi signal arises from the dynamic reflection pattern and can be produced from a wide range of movements, extending the range of potential applications.  Our technology uses machine learning to train the computer to identify each movement and could be applied to a variety of small movements. Currently, it has been used to determine the keystrokes of the user with high accuracy.  Applications extend from key stoke recording to biometric security and user identification.

Key Benefits

• Can function off of any commercial Wi-Fi device which offers CSI-waveforms.
• Sensitive to small disturbances.

Applications

• Human Computer Interaction (HCI) – use of gestures of the hand or fingers to control your computer.
• Biometrics – use of typing patterns to identify the user.
• Environments with small disturbances such as neonatal monitoring for breathing.

Patent Status: 

Software copyright

Licensing Rights Available

Full licensing rights available

Inventors: Kamran Ali, Alex X. Liu, Wei Wang, Nanjing University.

Tech ID: TEC2016-0023