Loading...
Please wait, while we are loading the content...
Similar Documents
Defenses against vibration-based covert channels in smartphones
| Content Provider | Indraprastha Institute of Information Technology, Delhi |
|---|---|
| Author | Khurana, Rushil |
| Abstract | In recent decades, the power of a mobile device has changed ten folds. With recent developments in the eld, the sensors present in the mobile devices are becoming extremely sophisticated. Unfortunately, access to some of these devices is unrestricted i.e. a malicious application can use this sensors to record information without the user coming to know about it. They don't require an explicit permission from the user to get installed. The accelerometer sensor in modern mobile devices has fairly high sensitivity. After circumventing weak access control systems, if any, a malicious application on a smartphone can analyze incoming accelerometer signals to covertly gain an understanding of activities in its physical surroundings in an unauthorized manner. By analyzing accelerometer readings, Marquardt et al. [1] showed that it was possible to covertly access the keystrokes typed on a PC keyboard using a mobile phone placed nearby. In this work, we replicate their work and examine defenses against such attacks that use accelerometer covert channels. We were able to show that while matching of abstracted words against context aware dictionary, we were able to recover as high as 84% of the words. We found that despite the hype, it is fairly easy to defend against this attack | inducing random noise into the machine learning technique leads to an immediate increase in privacy. Further improvements can be achieved by strategically induced noise; adding signals from acoustic noise such as playing a piece of music reduces the attack e ciency by more than 50%. |
| File Format | |
| Language | English |
| Access Restriction | Authorized |
| Subject Keyword | Security Mobile Computing Information Leakage Accelerometer |
| Content Type | Text |
| Educational Degree | Bachelor of Technology (B.Tech.) |
| Resource Type | Thesis |
| Subject | Data processing & computer science |
