Compromising location privacy through Wi-Fi RSSI tracking.

Area of Science:

• Computer Science
• Cybersecurity
• Mobile Computing

Background:

• Ubiquitous Wi-Fi and mobile devices generate vast amounts of data, raising privacy concerns.
• Existing limitations restrict access to Wi-Fi connectivity data like BSSID due to location inference risks.
• Alternative wireless data sources, such as RSSI, remain accessible and potentially exploitable.

Purpose of the Study:

• To investigate the feasibility of using Wi-Fi Received Signal Strength Indicator (RSSI) for inferring user presence and location.
• To assess the potential for correlating RSSI data with other features (e.g., time of day) to derive semantic context.
• To analyze the privacy implications and fingerprinting risks associated with accessible RSSI measurements.

Main Methods:

• Collected and analyzed Wi-Fi RSSI measurements from mobile devices in various common locations.
• Developed predictive models correlating RSSI data and temporal features (hour of day) to infer user presence.
• Evaluated the performance of these models in predicting user location and semantic context.

Main Results:

• Demonstrated that Wi-Fi RSSI can effectively infer user presence at specific locations over time.
• Achieved prediction performance exceeding 90% when correlating RSSI with contextual features like the hour of the day.
• Identified significant privacy risks and a substantial fingerprinting risk due to the unprotected nature of RSSI data.

Conclusions:

• Wi-Fi RSSI serves as a viable, yet unprotected, source for inferring user location and activity context.
• The findings underscore the critical need for enhanced privacy measures concerning the collection and use of RSSI data.
• Further research is warranted to develop robust defenses against RSSI-based location tracking and fingerprinting.