Short Bytes: Can your Wi-Fi router hear your talks or see what you do? It does sound like some crazy nightmare or an awesome future tech. The development of such human identification technologies has already begun and we may end up using such product in a matter of years.
A paper published by the researchers at the Northwestern Polytechnical University of China enables a modified Wi-Fi router to actually “see” a person present in the room. It does sound like a moonshot but a lot of research already been done in this field. Titled as FreeSense, their research enables the movements, shape, and size of a person.
A Wi-Fi router transmits radio waves to create a WLAN which is used by your device, such as a smartphone or tablet, to connect to the internet or any other device present on the same network. These radio waves are transmitted in 360 degrees covering almost every point in your room. A person present in the room interferes with the radio waves by obstructing their path. The router can analyze the obstruction patterns to get an idea of the person’s walking pattern and what he is trying to do.
This can be done even from behind a wall in some cases, like the research by MIT researchers Fadel Abib and Dina Katabi in 2013. It allowed a Wi-Fi router to sense hand gestures of a person on the other side of the wall by “capturing the reflections of its own transmitted signals off moving objects behind a wall in order to track them”. It can successfully detect if a person tries to play air guitar or draw some shape in the air.
The precision level could go all the way to detection of mouth movement to know what a person is saying. A research WiHear has been done to use WiFi signals to “hear” your talks. This involves analyzing the signal distortions of the WiFi signals occurred due to the movement of your mouth.
“We propose a novel approach for human identification, which leverages Wi-Fi signals to enable non-intrusive human identification in domestic environments,” the research paper reads. “It is based on the observation that each person has specific influence patterns to the surrounding Wi-Fi signal while moving indoors, regarding their body shape characteristics and motion patterns.”
Now, the system can’t detect a new person in the first shot. It requires some serious level of training before it manages to analyze a person in its vicinity. For example, for the system to detect one out of two people, it should be training with their physical presence in advance. Currently, FreeSence attains 94.5 percent accuracy when two people are present in the room and 88.9 percent accuracy in the case of six people.
The requirement of training could be thought of as a disadvantage to the practical usability of a system. But if optimistically considered, it can be beneficial. The system could easily spot an intruder because it doesn’t know much about him/her. Less amount of training can also aid to the privacy as the system won’t be able to circumvent the level of privacy.
The physical range of the Wi-Fi networks is also an important point to consider. It limits the operating area which could be used by the system for human identification. Possibilities are there that the system could be compromised a hacker.
FreeSense could also propel the development of supportive technologies for visually impaired people. The router can behave as their ears and eyes. Another application of FreeSense could be in the smart homes. When clubbed with the IoT ecosystem, it could set the lights and other stuff according to the preference of the person present in the room. However, it would be a difficult task to give preference to a family member. I can easily recall those brother-sister fights from my childhood. Technologies like these are interesting and sound promising, but still, a lot of development needs to be done before an end-product is pushed in front of the consumers.
If you have something to add, tell us in the comments below.