Changzhi Li received a National Science Foundation grant to figure out how multiple Wi-Fi-enabled devices can "share" the internet in a given space and leverage passive sensing to analyze human behaviors.
Since the onset of the COVID-19 pandemic, people have been getting used to the new normal of "virtual" everything – meetings, classes, games, you name it. While having a stable and fast internet connection is necessary, the extra devices in the household – computers, tablets, cellphones, etc. – have gobbled up the bandwidth, causing poor internet connections and slower speeds.
Changzhi Li, a professor of electrical and computer engineering in Texas Tech University's Edward E. Whitacre Jr. College of Engineering, received a $247,719 grant from the National Science Foundation (NSF) to develop a passive sensor to both reduce bandwidth competition between devices and help analyze human behaviors, such as vital signs. The three-year project officially began on Sept. 1.
"As the pandemic started, we began working intensively at home," Li said. "Myself, as an example, I had to teach virtual classes, and I also had to attend all kinds of virtual meetings. Sometimes, I would be kicked out of a meeting because of internet trouble. I upgraded my internet service to one gigabyte of bandwidth. However, it turns out that didn't completely solve the problem. So, I did some testing and figured out the problem is not the speed from our service provider, because one gigabyte is already fast enough. The true problem is the radio spectrum resource at home because we have multiple devices fighting for the limited wireless bandwidth available from the router, our access point."
The radio spectrum is a part of the electromagnetic spectrum. For Wi-Fi and Bluetooth communications, only very limited physical bandwidth is available because the electromagnetic spectrum must be appropriately shared among all wireless systems and applications, including GPS, garage door openers, wireless computer mice, keyboards and headsets, headphones, radio receivers, satellite television, broadcast television and radars, just to name a few.
"I thought, 'Can we reduce this type of competition among wireless devices?'" Li said. "Based on my research, I looked into sensing because it's becoming popular today to use wireless waves to sense various parameters around us. But most of the wireless sensors send out a wireless signal first, and the sensors analyze how the signal is modified by the subject of interest. From there, they extract information. So, the problem is, once we add wireless sensors to our real life, we are going to actually exaggerate this type of competition among wireless devices."
Li said a lack of radio spectrum resources is the one of the main reasons why a device suddenly loses its internet connection. However, Li said devices that use passive sensing, rather than active, could help reduce the number of signals competing for that limited bandwidth. An example of active sensing is when a police officer uses a radar gun to track how fast a car is going.
"The radar gun sends out a wireless signal toward our vehicles, and once that signal hits the vehicle, it bounces back," Li explained. "So, the radar gun can analyze and compare the transmitted signal with the reflected signal. From there, it can calculate the Doppler frequency shift, and from that, it figures out how fast a car is moving. But the problem of this is that, to sense the car in front of the police officer, first, the device has to send out a signal.
"Likewise, radars are widely used by today's cars for lane change assistance, collision avoidance, and parking assistance. This kind of extensive use of radio signals may create congestion to the wireless infrastructure. If everyone is using wireless sensors for health care and smart control at home, those wireless signals may conflict with each other. They are going to fight for bandwidth and for the time to be active. That, actually, is the reason why wireless devices get kicked out of the internet connection, simply because there are not enough radio spectrum resources to be shared among so many devices."
Li wants to develop a battery-operated device that will utilize injection-locking detection, which means the device will "lock" onto existing Wi-Fi signals in a living environment for sensing purposes. For the first two years of the three-year project, Li and his collaborators from Washington State University will develop the device's hardware and algorithm.
The hope is that the device Li and his cohorts are working toward can differentiate between a Wi-Fi signal coming directly from the router and Wi-Fi signals that are modified by human activities.
"By leveraging injection locking to sense how a subject alters or changes Wi-Fi signal in our environment, we can pick up useful information," Li said. "We can potentially use this for health care purposes. For example, we could measure vital signs, including heartbeats and respiration, 24/7. We could do this while the user is having breakfast or while the person is watching TV in the morning or at night.
"We also could do continuous authentication of users based on their behavior. For example, individuals have different ways of walking. The way they walk could change how the Wi-Fi signal propagates within the room. So, potentially, we could analyze that impact and identify if this person is authorized to be in the room."
