Investigators hacked a robotic vacuum cleaner to record speech and music remotely.
A team of researchers has shown that popular robotic vacuum cleaners in the home can be hacked remotely to act as microphones.
The researchers – including Nirupam Roy, an assistant professor in the Department of Computer Science at the University of Maryland – collected information from a laser-based navigation system on a popular vacuum robot and applied signal processing and applied in-depth learning techniques to identify speech and TV programs. in the same room as the device.
Research has shown that even without a microphone, any device that uses distance (Lidar) technology can be manipulated to detect sound. This work in collaboration with Assistant Professor Jun Han of the University of Singapore was presented at the Association for Computing Machinery’s Conference on Embedded Networked Sensor System (SenSys 2020) on 18 November 2020.
“We welcome these devices to our homes, and we don’t think about anything,” said Roy, who has a joint appointment at the Institute for Advanced Computer Studies at the University of Maryland (UMIACS). “But we’ve shown that even though these devices don’t have microphones, we can reuse the systems they use for browsing to spy on conversations and reveal private information.”
In the vacuum of the home, the bot’s Lidar navigation systems shine a laser beam around a room and detect the reflection of the laser when it bounces off nearby objects. The robot uses reflected signals to map the room and prevent collisions when moving from home.
Privacy experts have suggested that vacuum-mapped maps, often stored in the cloud, can lead to privacy breaches, introduce information to advertisers about home-size issues, suggest income levels and other lifestyle-related information. Roy and his team wondered if the Lidar of these robots could also pose security risks as a device for recording sound in users ’homes or businesses.
Sound waves cause objects to vibrate, and these vibrations cause small variations in the light that bounce off an object. Laser microphones, which have been used in espionage since the 1940s, are capable of converting these variations into sound waves. But laser microphones are based on a laser beam that is reflected from very smooth surfaces, such as glass windows.
An empty lid, on the other hand, scans the environment with a laser and senses light scattered back by objects of irregular shape and density. The scattered signal received by the vacuum sensor provides only a portion of the information needed to recover sound waves. The researchers were not sure if the empty Lidar system of a bot could function as a microphone and the signal could be interpreted into meaningful sound signals.
First, the researchers hacked an empty robot to control the position of the laser beam and show that the detected data can be sent to laptops via Wi-Fi without interfering with the device’s navigation.
They then conducted experiments with two sound sources. One source was a human voice reciting numbers played with computer speakers and the other was audio from multiple TV shows played through a TV soundbar. Roy and his colleagues captured the laser signal detected by the vacuum navigation system while it bounced off various objects located next to the sound source. Among the items are a trash can, a cardboard box, a carry-on bin and a polypropylene bag, which can usually be found on a typical floor.
The researchers were trained to learn the signals they received through in-depth algorithms to match human voices or to identify music sequences from TV shows. A computer system called LidarPhone has identified and matched oral numbers with 90%. accuracy. He identified the TV shows with a recording of more than 90% accuracy.
“This kind of threat may be more important now than ever, if we consider that we are all asking for food over the phone and holding meetings over the computer, and we are often talking about credit card or bank information,” Royk said. “But it’s even more worrying for me to reveal so much more personal information. This kind of information can tell you other things about my lifestyle, how many hours I’ve been working, what I’m doing. And what we see on TV can reveal our political orientation. That’s key for anyone who wants to manipulate political elections or direct very specific messages to me. ”
Researchers have pointed out that vacuum cleaners are just one example of the potential vulnerability to Lidar-based spying. Many other devices can be open to similar attacks, such as the infrared sensors in the phone used to recognize the face, or the passive infrared sensors used to detect movement.
“I think it’s a significant job to make manufacturers aware of these opportunities and provide solutions to the security and privacy community to prevent these types of attacks,” Royk said.
This research was partially funded by a grant from the Singapore Ministry of Education for Level 1 of the Academic Research Fund (R-252-000-A26-133 Award).
The research work, Spying with Your Robot Vacuum Cleaner: Eavesdropping via Lidar Sensors, Sriram Sami, Yimin Dai, Sean Rui Xiang Tan, Nirupam Roy and Jun Han, was presented on November 18, 2020, at the SenSys 2020 Computer Machinery Association. .
Reference: Sriram Sami, Yimin Dai, Sean Rui Xiang Tan, Nirupam Roy and Jun Han, November 2020, “Spying on your robot vacuum cleaner: listening through lidar sensors”. ACM SenSys 2020.
DOI: 10.1145 / 3384419.3430781