What is an ultrasonic frequency

Ultrasonic frequency: eavesdropping in high tones

In a study by the TU Braunschweig, more than 230 smartphone apps were identified that can use a special tactic to spy on their users: They listen for audio signals in the ultrasound range in order to determine the location of the mobile phone or on other devices owned by the user to be able to close.

The apps discovered with these "additional functions" are, for example, messenger services or service apps of large fast-food chains and some of them have been downloaded millions of times. They are all common in Southeast Asia.

Matthias Zeppelzauer from the Institute for Creative \ Media / Technologies at the St. Pölten University of Applied Sciences has been dealing with this type of audio tracking, which can bypass previous security devices on cell phones, for several years.

Together with colleagues, the IT researcher has now not only developed a method to discover and block the unwanted signals. The team is also working on using the technology in a more meaningful way - for example as a cheap form of communication between sensors and machines in industrial plants.

Inaudible to people

The transmission technology uses the ultrasonic range in the frequency band between 18 and 22 kilohertz, says Zeppelzauer. "This is an area in which humans can barely perceive sounds. Loudspeakers can, however, emit signals at this height, and most telephones can also receive them well through their microphones."

If, for example, this high-frequency noise is emitted via a TV set, it can be recorded by devices within a radius of up to ten meters. The mobile phone that receives the signals now transmits a corresponding message to the information collector's server.

This principle opens up a number of possibilities for data-hungry companies. "You can determine which devices are in a room and assign them to a user. Or you can identify users who interact with each other and thus record social networks," Zeppelzauer gives examples.

There are now a number of these ultrasonic technologies on the market. In the "SoniControl" project, supported by the Netidee initiative of the non-profit Internet Foundation Austria (IPA), the FH researchers have now developed an app that detects and blocks these signals in real time.

"The ultrasonic signals used are very different and have different types of modulation. We have therefore written an application that generally recognizes communication in this frequency band," explains Zeppelzauer. False detections caused by other noises in the ultrasonic range are filtered out.

Noise signal as a countermeasure

After a one-time authorization by the user, apps have access to the microphone of a cell phone at any time. The researchers therefore decided to develop a blocking function that disrupts the signal before it even reaches the smartphone.

"The ultrasound messages take about one to two seconds. That gives enough time to recognize the message and take countermeasures," emphasizes Zeppelzauer. This happens because the mobile phone itself actively emits ultrasonic tones that neutralize the incoming signal.

In future, the app, which is already available for download, will also recognize which ultrasound technology was used in the eavesdropping. In addition, a "map of ultrasound transmitters" could be created by sharing the data.

At the same time, the researchers are also thinking about applications other than tapping user data. After all, this type of signal transmission requires little energy and uses very inexpensive hardware with loudspeakers and microphones. "Everything you do with the Near Field Communication (NFC) radio standard can also be done with ultrasound," says Zeppelzauer.

In the "SonicTalk" project, a freely accessible open source protocol is to be created at the FH, with which developers can implement their respective technological solutions. "The ultrasound signals could be used well to monitor objects on a conveyor belt without having to resort to more expensive radio technology," the researcher gives an example. (Alois Pumhösel, June 16, 2018)