Researchers demonstrate “sonic gun” threat against smart devices
At the Black Hat security conference on Thursday, a team of researchers from Alibaba Security demonstrated how sound and ultrasound could be used to attack devices that depend on sensor input from gyroscopes, accelerometers, and other microelectromechanical systems (MEMS). A sonic "gun" could in theory be used to knock drones out of the sky, cause robots to fail, disorient virtual or augmented reality software, and even knock people off their "hoverboard" scooters. It could also potentially be used to attack self-driving cars or confuse air bag sensors in automobiles.
Many of the commercial gyroscope sensors in electronic devices are tuning fork gyroscopes—MEMS devices that use the vibrations of two "proof masses" to track rotation and velocity. But an outside source of vibration matching the resonant frequency of the gyroscope could interfere with the sensor's stability and cause the sensor to send bad data to the device it is embedded in.
By tuning an audio signal to the resonant frequency of the MEMS sensors, the researchers showed that the sensors could be temporarily disabled or manipulated into giving faulty data. In the process, this can cause the devices and software that depend on them to fail.
Wang Zhengbo and Wang Kang of Alibaba's mobile security research team demonstrated the threat with a home-built ultrasound emitting system. In an on-stage demonstration, they showed how sensors on Apple's iPhone 7 and the Samsung Galaxy S7 responded to the threat. The team also showed video of sonic attacks against commercial virtual reality headsets, a DJI drone, a Xiaomi "hoverboard," and a self-balancing robot toy (an attack that "makes some kids cry, Wang Kang joked).
Attacks on the hoverboard scooter with the improvised "sonic gun" required the emitter for the ultrasound to be either inside its outer shell or pressed to a seam in the shell. That's because the thick plastic reduced the strength of the sound waves able to reach the MEMS chips that provide balance sensing to the scooter's firmware. However, a higher-power emitter could theoretically penetrate the shell.
The researchers are continuing to investigate potential threats to other systems that rely on MIMS chips, including auto safety systems. Wang Kang said that much of the threat from sonic and ultrasonic jamming of sensors could be prevented by the use of sound-reducing or reflecting material in the shell of devices or by using software that actively monitors audio sensors and provides sound cancellation for frequencies that pose a threat. But the best fix long-term is the development of new types of MEMS sensors that are resistant to sonic attacks— such as "environmentally robust" gyroscopes that are less susceptible to external vibration.
MTCHT
ICT
TECHNOLOGICAL INNOVATIONS
POST
ABOUT US
NEWS
INTERESTING
INTERVIEW
ANALYSIS
ONLAIN LESSONS
