Today’s consumers trust their mobile phones with more personal information and data than ever. Consequently, when a device goes missing, it’s more than just an inconvenience—it’s a risk to their privacy and identity. Now, researchers have created an anti-theft system to better guard mobile devices and alert users of attempted thefts in real time.
While facial recognition and remote wiping systems keep strangers from gaining access to another person’s phone, little work has been done to prevent device theft in the first place. Researchers have developed a new software, called Virtual Safe, to track the unique walking pattern of the device’s owner and notify the owner when someone else has their device.
“Most anti-theft mechanisms for smart devices require user interaction,” said Dr. Dakun Shen of the research team. “These actions have to be triggered by the owner after they realize their device has been lost or stolen, so we wanted to design a new method that could automatically detect theft. Since thieves have to either walk or run away with a stolen device, we decided to use gait patterns to distinguish different users.”
Virtual Safe begins its theft detection process whenever a device is moved. It uses the phone’s accelerometer to monitor movement and immediately begins comparing the acceleration data of the phone’s owner to that of the person holding the device. Several other movement patterns, like step cycles, are then analyzed to determine if the current user is the device’s owner. If Virtual Safe believes a theft is in progress, an alarm is sent to the owner through a smart watch or email. An overview of the system architecture can be seen below in Figure 1.
Figure 1: Virtual Safe System Architecture
To ensure Virtual Safe was a viable option, the research team equipped 45 volunteers with its theft detection capabilities and tested its accuracy through a variety of movements, positions, and locations. They found Virtual Safe successfully detected unauthorized movement within 10-20 steps with a detection accuracy of 96.4%-98.4% and successfully distinguished movement as the owner 97.8% of the time.
Dr. Shen said his team now believes they have developed something that could be used beyond cell phones for theft detection. “Any other smart devices could apply this scheme to detect a stealing behavior, as long as they have the necessary sensors and basic computing power.”
For more information on mobile security, visit the IEEE Xplore digital library.