A Trajectory Released Scheme for the Internet of Vehicles Based on Differential Privacy

Abstract

The locations and users' information can be shared and interacted in the IoV (Internet of Vehicles), which provides sufficient data for traffic deployment and behavior pattern analysis. However, privacy issues had become more severe since personal or sensitive information is inclined to be revealed in a big data environment. In this work, a novel differential privacybased algorithm named DPTD (Differentially Private Trajectory Database) is proposed for trajectory database releasing. Firstly, a 3-dimensional generalized trajectory dataset is established by considering the time factor. Then, the trajectory space is divided into several planes through the timestamps, and the set of the locations on each plane is further processed by clustering and generalizing to re-form new trajectories, that is, the trajectories to be released. This method is quite favorable to prefix-tree releasing because the spatiotemporal characteristics of the trajectories can be captured and spareness problem is fixed. Besides, a Markov assumption-based prediction method is suggested in order to reduce the cost of adding noise. Unlike the traditional method that the noise is added layer by layer, the noise is only added to the odd layers based on the prediction through spatiotemporal correlation, saving approximately 50% of the privacy budget. Theoretical analysis and experimental results show that the proposed algorithm has better data availability than the compared algorithms while guaranteeing the expected privacy level.

SoK: Differentially Private Publication of Trajectory Data

“…A popular way [11,21,26,55,58,72,76,97,129] to quantify close data preservation is by using similarity measures, which output a value representing how different two trajectories are. For example, in mechanisms such that a one-to-one correspondence between the original and sanitized trajectories exists, we can use similarity measures to compute the average values between each pair.…”

“…Statistics preservation: In contrast with the previous categories, this one does not look at the preservation of the data comprising the database, but at specific extractable information. These statistics are extracted using query functions, and therefore the relative error query function [1,11,17,18,31,116,128] is frequently employed to study their preservation. Given the query q, it computes the difference between the outputs when using the original database D and the sanitized D ′ as…”

“…Some of the most common ones relate to visitor numbers and location popularity. For example, frequent sequential pattern mining looks at the k most common subtrajectories in the original and sanitized database, either by seeing if they match over the databases [17,18,127] or by comparing the counts of such [11,17,114]. Similarly, count queries [17,18,76,128] can be utilized to check whether the number of visitors to locations is retained or not.…”

“…Similarly, count queries [17,18,76,128] can be utilized to check whether the number of visitors to locations is retained or not. Additionally, some metrics tackle the preservation of number of trajectories [76], most visited locations [78], hotspots [26], location popularity [114], flow density [11,114] and the start and end points distribution (trip error [114]).…”

“…The next category contains mechanisms [11,21,55,58,72,129] that cluster locations and subsequently release trajectories through these clusters with some perturbation to guarantee privacy.…”