Publishing Sensitive Trajectory Data under Enhanced l-Diversity Model

With the proliferation of location-aware devices, trajectory data have been widely collected, published, and analyzed in real-life applications. However, published trajectory data often contain sensitive attributes, so an attacker who can identify an individual from such data through record linkage, attribute linkage, or similarity attacks can gain sensitive information about this individual. To resist from these attacks, we propose a scheme called Data Privacy Preservation with Perturbation (DPPP). To protect the privacy of sensitive information, we first determine those critical location sequences that can identify specific individuals. Then we perturb these sequences by adding or deleting some moving points while ensuring the published data satisfy (1, alpha, beta)-privacy, an enhanced privacy model from l-diversity. Our experiments on both synthetic and real-life datasets suggest that DPPP achieves better privacy while still ensuring high utility, compared with existing privacy preservation schemes on trajectory.