Abstract: As more people carry smartphones and portable devices, people-centric sensing using these devices can provide high quality information about the environment, the society, and the individual human carriers. In densely populated areas, sensor-based applications can benefit from sensor sharing between neighboring wireless devices. With sensor sharing, a device can collect sensor data that it cannot collect using its own sensors. Even if all necessary sensors are available, collaborating with other devices can help optimize energy consumption and data quality due to the diversity in their efficiency and sensor accuracy.
Sensor sharing, however, poses many threats to the privacy of carriers. To share sensors, each device has to reveal what sensors they have and what sensors they lack. As people carry more sensors differing in their modality, quality, and energy efficiency, the sensor-inventory information can be used to identify the carrier. We propose AnonyShare to address privacy concerns raised by sensor sharing; AnonyShare allows carriers to anonymously share their sensors. To mitigate inventory-based attacks, we quantify the privacy risk of sensor sharing and provide a sharing policy that tries to keep the privacy risk as low as desired. We evaluate the feasibility of AnonyShare by implementing it with off-the-shelf mobile devices.
No online copy available.
Copyright © 2009 by IEEE.