Abstract:

Vehicles carrying GPS equipment create large trajectory information. Analyzing and publishing trajectory data flow statistics based on road network is beneficial to the improvement of network structure and the realization of intelligent transportation. However, the direct release of trajectory traffic can lead to the disclosure of user privacy, and there is lack of a rigorous and provable privacy method to release traffic flow in road networks. Therefore, this paper presents a differential privacy trajectory flow releasing method. The method is divided into two steps: firstly, the flow value of each section is statistically calculated and the difference privacy noise is added. Secondly the post adjustment algorithm is put forward for the consistency characteristic of the flow graph, so that the adjusted flow graph not only satisfies the consistency characteristic, but also greatly reduces the publishing error. The experiment on the real road network shows that the method?has the ability to deal with large-scale network traffic, and after the optimization of the post adjustment algorithm, the release error is reduced by about 13%.

Key words: trajectory flow, differential privacy, consistency adjustment, road network constraints

摘要：

搭载GPS设备的车辆在运行过程中产生大量轨迹信息，对轨迹流量信息的统计与发布有利于改善路网结构，实现智能交通。但是直接发布轨迹流量可能导致用户隐私的泄露，而目前缺乏严格的可证明的轨迹流量隐私保护发布方法。为此，提出了一种基于路网的差分隐私轨迹流量发布方法。该方法分两步：首先根据轨迹数据统计各个路段的流量值并添加差分隐私噪声；随后针对流量图的一致性特性提出后置调节算法，使得调节后的流量图不仅重新满足一致性特性，而且还极大地减少了发布误差。在真实路网上的实验表明，该方法具有处理大规模路网流量的能力，且经过后置调节算法的优化，发布误差减小了约13%。

关键词: 轨迹流量, 差分隐私, 一致性调节, 路网约束