WiSee Enables WiFi Gesture Recognition - University of Washington Research
@satya-swaroop-YDeBJM
•
Oct 23, 2024
Oct 23, 2024
1.6K
Last year on VoiCE we had brought you news on how University College, London has managed to use the Doppler Effect to <a href="https://www.crazyengineers.com/threads/wi-fi-router-enables-you-to-see-through-walls.66279">Wi-Fi Router Enables You To See Through Walls</a>. Taking this technology forward is a team of researchers from University of Washington led by Shwetak Patel, an assistant professor of computer science and engineering and electrical engineering who have managed to make gesture recognition technology possible by using only Wi-Fi signals. For this project they took a standard Wi-Fi router and modified it as a receiver so that it can ‘listen’ to all the Wi-Fi frequency range signals coming from gadgets like computers, smartphones and tablets. Next they applied the principle of Doppler Shift to detect the pattern of changes that occur when a person moves his hands or legs. The team developed an algorithm that would detect these changes of signal and also made it sure that they account for gaps in wireless signals that are made when a Wi-Fi device is switched off.
WiSee, which is in its prototype stage can identify nine different whole-body gestures, ranging from pushing, pulling and punching to full-body bowling. The researchers have tested these gestures with 5 users and have found that WiSee can detect these gestures with an accuracy of 94%. The system uses multiple antennas to focus only on a single user’s gestures, this makes sure that the system does not get confused by other people present in the room who might be inadvertently doing the same gestures. Since Wi-Fi works across rooms this technology makes sure that you are able to send commands to your gadgets even if you are in the next room or even outside your house. To make sure that the above feature is not exploited by hackers a specific repetition gesture sequence has been developed that acts as a password for accessing the WiSee system. See this technology at work in the video below:
WiSee, which is in its prototype stage can identify nine different whole-body gestures, ranging from pushing, pulling and punching to full-body bowling. The researchers have tested these gestures with 5 users and have found that WiSee can detect these gestures with an accuracy of 94%. The system uses multiple antennas to focus only on a single user’s gestures, this makes sure that the system does not get confused by other people present in the room who might be inadvertently doing the same gestures. Since Wi-Fi works across rooms this technology makes sure that you are able to send commands to your gadgets even if you are in the next room or even outside your house. To make sure that the above feature is not exploited by hackers a specific repetition gesture sequence has been developed that acts as a password for accessing the WiSee system. See this technology at work in the video below:
For more details head over to <a href="https://www.washington.edu/news/2013/06/04/wi-fi-signals-enable-gesture-recognition-throughout-entire-home/" target="_blank" rel="nofollow noopener noreferrer">Wi-Fi signals enable gesture recognition throughout entire home | UW News</a> or read their report <a href="https://wisee.cs.washington.edu/wisee_paper.pdf" target="_blank" rel="nofollow noopener noreferrer">PDF</a> [PDF file]