"I fell asleep on my pillow at night, but I didn't sleep alone."
The Syrian poet Adonis did not know how many insomniacs were mentioned. Today, when the pace of life is getting faster and faster, it is quite a luxury for many people to wake up to sleep naturally. According to a survey by the World Health Organization, 38% of people in China have different levels of sleep disorders.
Many people began to use the power of technology to understand their sleep, and a large number of smart bracelets came into being. But for some people, wearing a bracelet to sleep itself is a kind of interference, but now artificial intelligence is helping you solve this problem, not only hope to get rid of insomnia, but also get rid of those wearable devices.
The Massachusetts Institute of Technology recently published a new study that allows an AI algorithm to monitor sleep by analyzing radio waves, replacing the monitoring mode that users need to wear sensors. It is reported that this AI algorithm can analyze the user's pulse, breathing and other values ​​to identify whether the user is in several stages of mild, deep or rapid eye movement sleep (REM).
So since you don't have to wear any wearable device, how does the AI ​​algorithm get a series of data from our sleep? Dina Katabi, professor of electrical engineering and computer science at the Massachusetts Institute of Technology, gave the answer:
Imagine if your wireless router knows that you are dreaming, you can also identify if you are in deep sleep.
The research team led by Katabi hopes to develop a health sensor consisting of a wireless router that monitors the user's physiological signals and important health indicators at night, but does not require any changes from the user. .
According to Katabi, this sensor is similar to a smart Wi-Fi box, the size of a notebook computer, can emit low-power radio frequency (RF) signals, any slight movement of the human body will change the frequency of the reflected signal, the sensor is analyzed These frequency changes are used to obtain vital sign data such as pulse and respiratory rate.
Prior to this, Katabi and her students had developed a sensor called WiGait that uses wireless signals to measure walking speed and helps doctors diagnose health problems such as decreased cognitive ability and cardiopulmonary disease. Inspired by Katabi, Katabi decided to extend the reach of this sensor to sleep monitoring.
However, since the reflected signal is based on long-distance measurement, the sensor can easily receive information that is not related to sleep, and the monitoring result will be disturbed. In fact, even with built-in high-sensitivity non-wearing devices such as wearable devices such as Fitbit smart bracelets and smart pillows, these monitoring devices are more likely to be “spoofed†and the resulting data may not truly reflect our sleep.
At this time, it is necessary to rely on artificial intelligence. The research team combines three deep neural network AI algorithms to obtain more accurate data.
When the sensor receives the reflected signal, the first layer neural network will initially parse the data through image recognition, and then the second layer neural network is responsible for measuring which sleep stage the user is in, and the last layer of neural network will further perform the data. Analysis and comparison.
The research team has tested this monitoring program in the sleep of 25 volunteers, and the accuracy of returning this monitoring technology is as high as 80%. This is in line with the electroencephalogram used by sleep experts to fully cover the sensor device. The accuracy of the (EEG) measurement is comparable. Tommi Jaakkola, a member of the study, said:
This neural network algorithm can effectively identify the sleep signal and discharge other unrelated signal interference, and the algorithm does not need to be calibrated after the sensor position or the monitored object changes.
It can be seen that this AI-based sleep monitoring not only allows users to get rid of those complex wearable devices, but also reduces the difficulty of medical staff.
The research team is currently using this monitoring program to study how Parkinson's disease affects sleep, and the sensor can help researchers further understand sleep disorders such as insomnia and sleep apnea.
It can be imagined that if this artificial intelligence wireless monitoring technology continues to advance or expand the application scenario, and thus become the standard for every smart phone, perhaps your Siri will really become the intelligent AI housekeeper of Iron Man's Jarvis.
Active Stylus Pen,Stylus Pencil,Capacitive Stylus Pen,Tablet Pencil
Shenzhen Ruidian Technology CO., Ltd , https://www.wisonens.com