Crowd localization: Threats and solutions

The use of Bluetooth Low Energy (BLE) recently became famous in crowd localization services. In this article we show that such service induces significant threats and we propose a scheme that achieves privacy preserving crowd localization. Crowd localization applications leverage all their users to perform periodic BLE scan and detect the nearby tracked devices. The users run a smartphone application which periodically forwards the list of tracked devices seen in the vicinity. This allows a user to track the mobility of its tracked devices at a large scale. We show that these tags use public or random static addresses that identify the user. These addresses are sent to the tracking company server without modifications. Such a behavior puts unnecessary risk on the users as its mobility can be tracked by the tracking company and by any malicious person with a BLE smartphone. We propose a new architecture that performs crowd localization while preserving users privacy. Our proposal prevent tracking of the users localization by the company as well as any BLE capable adversary. Compared to the current proposals it has no impact on the sensors or smartphone's battery life. It doesn't impair the performance of the crowd localization. Its only downside is a slight increase of the computation cost for the tracking company servers when a user declare the loss of its device.