Traffic density detection is a useful measure to infer the environment of road condition. The common methods to achieve traffic density detection is collecting the real time vehicle locations in the areas. But processing data that consist of vehicle locations severely risks car owner's privacy.

In this paper, we present a vehicle location privacy protection framework called Expanding Geo-Indistinguishability framework (EGeoIndis). By abstracting map as bitmap, we utilize Linear Programming (LP) to achieve geo-indistinguishability protection in low quality loss. We creatively combine a secret sharing protocol to resolve computational efficiency problem when we meet large-scale locations in LP and finally we achieve a better privacy guarantee than geo-indistinguishability in large areas. All our privacy data achieve the protection locally, so we don't need any third trusted party. Evaluation result shows that EGeoIndis has a better performance than classic Laplace mechanism when setting proper parameter in availability and privacy guarantee meanwhile has a good efficiency.

交通密度检测是推断路况环境的有用措施。实现交通密度检测的常用方法是收集区域中的实时车辆位置。但是处理包含车辆位置的数据会严重威胁车主的隐私。

在本文中，我们提出了一种称为“扩展地理不可区分性框架（EGeoIndis）”的车辆位置隐私保护框架。通过将地图抽象为位图，我们利用线性编程（LP）来实现低质量损失的地理不可区分性保护。当我们在LP中遇到大规模地点时，我们创造性地结合了秘密共享协议来解决计算效率问题，最后，与大范围的地理不可分辨性相比，我们获得了更好的隐私保证。我们所有的隐私数据均在本地获得保护，因此我们不需要任何第三方信任的人。评估结果表明，在可用性和隐私保证方面设置适当的参数时，EGeoIndis的性能优于经典的Laplace机制。