MIT researchers have developed a wireless, private method of monitoring a person’s sleeping posture – whether on their back, stomach, or sides – with reflected radio signals from a small device on the wall of a bedroom.
According to Shichao Yue, the device with the name BodyCompass is the first high-frequency-based system provided at home that provides accurate sleep data without cameras or sensors attached to the body. He will introduce the system in a presentation at the UbiComp 2020 conference on September 15th. The PhD student has been using wireless sensors to study sleep stages and insomnia for several years.
“We thought the sleep posture could be another effective application of our system,”
“Unfortunately, many patients have no idea how they sleep at night or what position they end up in after a seizure,” says Dong Woo Lee, an epilepsy neurologist at Brigham and Women’s Hospital and Harvard Medical School, who is unrelated to the study was brought. “A body monitoring system like BodyCompass would advance our field by enabling basic monitoring of our patients to assess their risk and, in combination with an alert / intervention system, could save patients from sudden, unexpected death in epilepsy.”
In the future, people could also use BodyCompass to track their own sleep habits or monitor the sleep of infants. Yue says, “It can be either a medical device or a consumer product, depending on your needs.”
Other authors of the conference paper published in the ACM practices on interactive, mobile, wearable and ubiquitous technologiesThese include PhD students Yuzhe Yang and Hao Wang and Hariharan Rahul, a subsidiary of Katabi Lab. Katabi is Andrew and Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT.
BodyCompass analyzes the reflection of radio signals when they bounce off objects in a room, including the human body. Similar to a WiFi router mounted on the wall of the bedroom, the device sends and collects these signals as they return via multiple paths. The researchers then map the paths of these signals and work backwards from the reflections to determine posture.
In order for this to work, however, the scientists had to find out which of the signals bounced off the sleeper’s body and not from the mattress, a bedside table or a ceiling fan. Yue and his colleagues realized that their previous work on deciphering breathing patterns from radio signals might solve the problem.
Signals bouncing off a person’s chest and stomach are uniquely modulated by breathing. Once this breathing signal was identified as a way to “mark” reflections from the body, the researchers were able to analyze these reflections versus the location of the device to determine how the person was lying in bed. (For example, if a person was lying on their back, strong radio waves bouncing off their chest were directed at the ceiling and then at the device on the wall.) “Recognizing breathing as coding helped us separate signals from the body and environmental reflections that we can use to keep track of where informative reflections are, ”says Yue.
Reflections from the body are then analyzed by an adapted neural network to infer how the body is angled while sleeping. Since the neural network defines sleeping postures according to angles, the device can distinguish between a sleeper lying on the right side and a sleeper who is only slightly inclined to the right. This type of fine-grained analysis would be especially important for epilepsy patients, where sleeping on the prone position correlates with sudden unexpected death, says Yue.
Lee says, “It is becoming apparent that patients don’t like to wear devices, they forget to wear them, they are less comfortable, battery life is short and data transfer can be difficult. A non-wearable contactless device like the BodyCompass would solve these problems. ” . “
BodyCompass offers several advantages over other methods of monitoring sleep posture, e.g. For example, installing cameras in a person’s bedroom or attaching sensors directly to the person or their bed. Sleeping with sensors can be uncomfortable and cameras compromise someone’s privacy, Yue notes. “Since we only record important information for recognizing sleep posture, such as a person’s breathing signal during sleep, it is almost impossible for anyone to infer other user activities from this data.”
An accurate compass
The research team tested the accuracy of BodyCompass over 200 hours of sleep data from 26 healthy people who sleep in their own bedrooms. At the beginning of the study, the subjects wore two accelerometers (sensors that detect movement) attached to their chest and stomach to train the device’s neural network with data on “ground truth” in their sleeping posture.
BodyCompass was most accurate – predicting correct posture 94 percent of the time – when the device was trained on a week’s worth of data. One night’s training data provided accurate results 87 percent of the time. BodyCompass was able to achieve 84 percent accuracy with just 16 minutes of data collected when the sleepers were asked to hold some common sleeping postures in front of the wireless sensor.
BodyCompass, along with epilepsy and Parkinson’s disease, could be useful in treating patients prone to pressure ulcers and sleep apnea, as both conditions can be relieved by changing sleep posture. Yue also has his own interest: he suffers from migraines, which apparently depends on how he sleeps. “I sleep on my right side to avoid a headache the next day,” he says, “but I’m not sure there is really a connection between sleeping posture and migraines. Maybe this can help me find out if there is a relationship. “
Currently, BodyCompass is a monitoring tool, but it may one day be combined with an alert that can prompt sleepers to change their posture. “Researchers are working on mattresses that can turn a patient slowly to avoid dangerous sleeping positions,” says Yue. “Future work could combine our sleep posture detector with such mattresses to bring an epilepsy patient into a safer position if necessary.”