An ‘AI’ Is Using ‘Ultrasonic Sound Waves’ To Track People’s Movement

ad1

As technology progresses, and sentiments against invasive techs grow, the tech community, together with the academe, has been developing new alternatives to existing systems that are more palatable to the people. One example of this technology is the latest AI trained in China that can detect and survey people just by using echolocation.

A Chinese researcher has found a way to balance security and privacy through artificial intelligence that can allow computers to echolocate. By training AI to sift through signals from arrays of acoustic sensors, the system can gradually learn to parse your movements—standing, sitting, falling—using only ultrasonic sound.

According to Dr. Xinhua Guo at the Wuhan University of Technology, the newly developed system can be more palatable to privacy advocates, as it is not as invasive as cameras or facial recognition software. As artificial intelligence relies on echolocation and ultrasonic sounds, similar to the technology that allows people to navigate in the dark, it does not invade privacy since it does not record any video or audio. This means that it only tracks motion and not the person per se.

“Human activity recognition is widely used in many fields, such as the monitoring of smart homes, fire detecting and rescuing, hospital patient management, etc. Acoustic waves are an effective method for human activity recognition,” Dr. Xinhua Guo wrote in his study.

“In traditional ways, one or a few ultrasonic sensors are used to receive signals, which require many feature quantities of extraction from the received data to improve recognition accuracy.”

The system uses a two-dimensional acoustic array and convolutional neural networks. A single feature quantity is utilized to characterize the sound of human activities and identify those activities.

Photo: Dr. Xinhua Guo

For those who are adamant about the accuracy of the system, the results of the tests conducted by the researcher show 97.5% for time-domain data and 100% for frequency-domain data. This level of accuracy surmounts the accuracy rate of using facial recognition software. Aside from facial recognition, the new system was said to outperform other motion tracking technologies like k-nearest neighbor and support vector machines.

However, Guo admits that the system can better identify static motions such as sitting or standing still rather than movements per se. This is expected, explained the authors, because falling and walking introduce individual differences in how people move, making it harder for computers to figure out a general acoustic pattern. But with future studies, the researcher aims to resolve this gap in his development.

It could potentially be used in many applications

As the new system could improve over time through further research, it could be used in a plethora of different applications from health, to home security, to law enforcement operations. The system could allow caretakers to monitor elderly folks who live alone for falls inside their house or track patient safety inside hospital rooms.

The possibility and the potential for this system are endless. It could be installed in public spaces to monitor motions in trains, sidewalks, and public utility vehicles to guard people against sexual harassment or violence similar to how satellites are used to capture heat images of the sea to patrol for sunken ships, maritime collision, and other nautical incidents.

However, since the system only records motion, there is no potential for identification – unlike facial recognition and surveillance cameras. A feature that experts believe could be the perfect balance between safety and privacy in technology. The system doesn’t even generate blob-like body shapes that grace US airports’ body camera screens.

The clamor against facial recognition

This could be good news for privacy advocates and people in general. Clamor has since been loud regarding the potential for privacy violations by facial recognition software. And the scientific community is in tune with advocates in saying that facial recognition technology is a juvenile technology that needs regulation as it tramples the rights of people to privacy.

Only recently, reports reveal that law enforcement agencies have been using facial recognition technologies to solve crimes – sometimes in creative ways. A police station has been feeding images of celebrities who they think could look like their suspect. And arrests have been made. But it seems like this strategy is only accurate less than half the time; which lead to incidents of wrongful detention in many cases.

Because of the intensifying opposition of the scientific community and people in general of facial recognition technology, San Francisco became the first city to outlaw the use of facial recognition techs in police operations and by other city agencies to serve as a model for other cities to follow through.

Leave a Reply

Your email address will not be published. Required fields are marked *