Never Walk Alone: Uncertainty for Anonymity in Moving Objects Databases

Abstract

Abstract-Preserving individual privacy when publishing data is a problem that is receiving increasing attention. According to the k-anonymity principle, each release of data must be such that each individual is indistinguishable from at least k − 1 other individuals. In this paper we study the problem of anonymity preserving data publishing in moving objects databases. We propose a novel concept of k-anonymity based on co-localization that exploits the inherent uncertainty of the moving object's whereabouts. Due to sampling and positioning systems (e.g., GPS) imprecision, the trajectory of a moving object is no longer a polyline in a three-dimensional space, instead it is a cylindrical volume, where its radius δ represents the possible location imprecision: we know that the trajectory of the moving object is within this cylinder, but we do not know exactly where. If another object moves within the same cylinder they are indistinguishable from each other. This leads to the definition of (k, δ)-anonymity for moving objects databases.We first characterize the (k, δ)-anonymity problem and discuss techniques to solve it. Then we focus on the most promising technique by the point of view of information preservation, namely space translation. We develop a suitable measure of the information distortion introduced by space translation, and we prove that the problem of achieving (k, δ)-anonymity by space translation with minimum distortion is NP-hard. Faced with the hardness of our problem we propose a greedy algorithm based on clustering and enhanced with ad hoc pre-processing and outlier removal techniques. The resulting method, named N WA (N ever Walk Alone), is empirically evaluated in terms of data quality and efficiency.Data quality is assessed both by means of objective measures of information distortion, and by comparing the results of the same spatio-temporal range queries executed on the original database and on the (k, δ)-anonymized one. Experimental results show that for a wide range of values of δ and k, the relative error introduced is kept low, confirming that N WA produces high quality (k, δ)-anonymized data.

Microaggregation- and permutation-based anonymization of movement data

“…Consequently, several anonymity notions and methods for trajectories have been proposed [21,20,22,7,37,6,1,33,43,31,34,47,32,2,23,24]. Among those works, we next review the ones that are most similar to our approach, and we highlight our comparative advantages.…”

“…We theoretically and experimentally compare our first heuristic with a recent trajectory anonymization method called (k, δ)-anonymity [1] also aimed at trajectory k-anonymity without reachability constraints. Theoretical results show that the privacy preservation of our first method is the same as that of (k, δ)-anonymity but dealing with trajectories not having the same time span.…”

“…k-Anonymity cannot be directly achieved with spatio-temporal data, because any point or time can be regarded as a quasi-identifier attribute [1]. Direct k-anonymization would require a set of original trajectories to be transformed into a set of anonymized trajectories such that each of the latter is identical to at least k − 1 other anonymized trajectories.…”

“…· The second method takes reachability constraints into account, and it tries to reduce the fraction of discarded locations by replacing trajectory k-anonymity with location k-diversity; • Empirical results on synthetic and real-life data portraying the performance of the two above methods. Both methods are confronted with (k, δ)-anonymity [1], which has some comparable features.…”

“…Closest to our approach is the notion of (k, δ)-anonymity [1,2]. In the original method -Never Walk Alone (NWA) [1]-, the set of trajectories is partitioned into disjoint subsets in which trajectories begin and end at roughly the same time; then trajectories within each set are clustered using the Euclidean distance.…”