A recent study from IIT Delhi has demonstrated that Android apps with precise location access can extract more user information than previously understood. Led by MTech student Soham Nag and Prof Smruti R Sarangi, the system named AndroCon uses GPS data to infer context and activities with high accuracy.
Methodology and Results
AndroCon utilizes nine GPS parameters including Doppler shift and multipath interference to detect user activities like sitting, standing, and traveling. The system also maps indoor spaces with an error margin under four meters. Conducted over a year with various smartphone models across a 40,000 sq km area, the study demonstrated up to 99% accuracy in identifying surroundings and more than 87% accuracy for human activities.
Implications for User Privacy
The study identifies a significant privacy issue: any Android app with location permissions can potentially collect sensitive data without user awareness. AndroCon's findings suggest that users may unknowingly share more personal information through their location data than intended. This research is detailed in the ACM Transactions on Sensor Networks.
