Jiayi
Chen,
PhD
candidate
David
R.
Cheriton
School
of
Computer
Science
Smartphone loss affects millions of users each year and causes significant monetary and data losses. Device tracking services (e.g., Google's "Find My Device") enable the device owner to secure or recover a lost device, but these services can be easily circumvented by the person who finds the device. An effective smartphone loss prevention solution should alert the owner before they leave the premises without the phone.
We present Chaperone, an open source, real-time smartphone loss prevention system that does not require additional hardware. Chaperone adopts active acoustic sensing to detect the user's departure via the built-in speaker and microphone. It is designed to robustly operate in real-world scenarios characterized by bursting high-frequency noise, bustling crowds, and diverse environmental layouts. We evaluate Chaperone by conducting over 1,100 experiments at a variety of locations including coffee shops, restaurants, offices, transit stations, and cars under different testing conditions.
Our evaluation shows that Chaperone provides an overall precision of 96.2% and an overall recall of 97.5% for device loss detection. Furthermore, Chaperone is able to detect smartphone loss and alert the device owner within one second for 95% of the successful detection cases. Finally, we provide an implementation of Chaperone as a standalone smartphone app and show that it has an acceptable battery consumption rate of 5.6% per hour during active acoustic sensing.