June 18, 2018
About a month ago, researchers from the University of Manchester and University of Madrid unveiled a new AI-Based gait recognition system, which can be used for conducting non-intrusive security scans. Now, researchers at MIT have developed a new project named ‘RF-Pose’ that uses AI and RF waves for similar and critical applications.
The research paper published by the team has also described the whole mechanism of the system, “While visible light is easily blocked by walls and opaque objects, radio frequency (RF) signals in the WiFi range can traverse such occlusions. Further, they reflect off the human body, providing an opportunity to track people through walls.”
The reflected waves then create a vague heatmap, which is then fed to a trained neural network to create stick figures of humans. These stick figures will then reflect all the movements made by the humans, or lack of, behind the walls.
Source: Through-Wall Human Pose Estimation Using Radio Signals report by MIT
It is clear to see how this system has the potential to be easily implemented for less-intrusive security scans by detecting the human movement without using cameras. Since this system is completely based on radio frequency waves, it can be used even in the darker environments for undeterred accuracy.
The team, however, wishes to use the application for healthcare purposes. The corresponding news report published by MIT, states that the system can be effectively used to monitor ailments like Parkinson’s disease, muscular dystrophy and multiple sclerosis. The team says that with this system, the doctors will have a better tool to detect the progression of the disease and adjust the medications accordingly.
Professor Dina Katabi, the head of the research, has stated, “We’ve seen that monitoring patients’ walking speed and ability to do basic activities on their own gives healthcare providers a window into their lives that they didn’t have before, which could be meaningful for a whole range of diseases. A key advantage of our approach is that patients do not have to wear sensors or remember to charge their devices.” Furthermore, the same capabilities also make the system fit for the elderly people who are living independently. It can be used to detect movements like falls, injuries, or even a slight change in activity pattern to trigger an alarm.
Protection of User Privacy
Non-intrusive or otherwise, it was important for the researchers to develop the system’s foundation to protect user privacy in later applications. So, all the data that was used to train the neural network was collected only after getting formal consent from the subjects. Furthermore, the data collected was also anonymised and encrypted for more security. For real-world applications of the system, researchers have plans to embed a ‘consent mechanism’ in it. Doing so will require the people to make a specific set of movements to cue the device to begin the monitoring of the environment.
Besides healthcare and security, the technology can also be used for designing new classes of games like Kinect. However, unlike the other games that work on the model that presents motion detection technologies, these games would be more advanced and can encompass wider areas like a whole house, without the cumbersome wiring requirements.
The report also mentions another critical application of the technology in search-and-rescue missions to help locate survivors. The demonstration of the technology has shown that it doesn’t only detect the movement but also identify individuals with an accuracy of 83%, making its use in the aforementioned missions more functional.