A mobility prediction system leveraging realtime location data streams

Kulkarni, Vaibhav; Moro, Arielle; Garbinato, Benoît

doi:10.1145/2973750.2985263

Cited by 10 publications

(20 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We use the same technique proposed in [3] to extract clusters and then derive the parameter values across two geospatial datasets as shown in Figure 9a. To select the value of minimum visit, either the knee in the plot of time vs. number of visits to a particular ROI is selected [21] as shown in Figure 9b, or the duration of the collected dataset is taken into account [9]. We clearly see different trends followed by the two datasets.…”

Section: The Parameter Cursementioning

confidence: 97%

“…A mobility model consists of three main components: (i) ROIs, (ii) representative paths, and (iii) transition probabilities. Once all the ROIs are computed, the process of obtaining the representative paths and the probabilities is detailed in our previous work [5], [21]. Constructing a representative path is essentially a procedure to efficiently extract the set of Cell IDs that best describes the trajectory between the two ROIs.…”

Section: Mobility Modelingmentioning

confidence: 99%

“…Constructing a representative path is essentially a procedure to efficiently extract the set of Cell IDs that best describes the trajectory between the two ROIs. The transition probabilities can be computed by using mobility Markov chains [21].…”

Section: Mobility Modelingmentioning

confidence: 99%

“…A mobility model is thus formulated by linking all the distinct ROI's, the representative paths and the transition probabilities. The transition probabilities are estimated based on a mobility Markov chain (MMC) model which accounts for the state-transition matrix as described in [21].…”

Section: Qs =mentioning

confidence: 99%

See 3 more Smart Citations

Capstone: Mobility Modeling on Smartphones to Achieve Privacy by Design

Kulkarni

Moro

Chapuis

et al. 2018

2018 17th IEEE International Conference on Trust, Security and Privacy in Computing and Communications/ 12th IEEE International

Self Cite

View full text Add to dashboard Cite

Sharing location traces with context-aware service providers has privacy implications. Location-privacy preserving mechanisms, such as obfuscation, anonymization and cryptographic primitives, have been shown to have impractical utility/privacy tradeoff. Another solution for enhancing user privacy is to minimize data sharing by executing the tasks conventionally carried out at the service providers' end on the users' smartphones. Although the data volume shared with the untrusted entities is significantly reduced, executing computationally demanding server-side tasks on resourceconstrained smartphones is often impracticable. To this end, we propose a novel perspective on lowering the computational complexity by treating spatiotemporal trajectories as space-time signals. Lowering the data dimensionality facilitates offloading the computational tasks onto the digital-signal processors and the usage of the non-blocking signal-processing pipelines. While focusing on the task of user mobility modeling, we achieve the following results in comparison to the state of the art techniques: (i) mobility models with precision and recall greater than 80%, (ii) reduction in computational complexity by a factor of 2.5, and (iii) reduction in power consumption by a factor of 0.5. Furthermore, our technique does not rely on users' behavioral parameters that usually result in privacy-leakage and conclusive bias in the existing techniques. Using three real-world mobility datasets, we demonstrate that our technique addresses these weaknesses while formulating accurate user mobility models.Problem 1: Mobility Signal Generation. Given a trajectory T u of an individual u, a temporally ordered sequence of tuples, 3 Art. 23: www.privacy-regulation.eu/en/article-23-restrictions-GDPR.htm such that, T u = (l 1 ,t 1 ), (l 2 ,t 2 )...(l n ,t n ) , where l i = (lat i , lon i ), the latitude-longitude coordinate pair and t, the timestamp such that t i+1 > t i , translate T u into a 2-D signal S u (t), modeled as a function of changing distance with respect to time.Requirements and Challenges. (i) Constructing a continuous graph from the noisy and non-uniformly sampled location trajectories, (ii) preserving all the key knowledge contained in the trajectory samples, and (iii) retaining the spatial locality between the discretized points.Problem 2: Signal Interpretation. Given a user's spatiotemporal signal S u (t), interpret and model the distinct signal elements in the temporal and spectral domain, i.e. local maxima/minima, rising/falling edges, static signal component, candidate frequencies, spectral coefficients and harmonics, with respect to human mobility behaviors.Requirement and Challenge. In order to facilitate interdomain switching, attach and validate a semantic meaning to each of the above signal components.Problem 3: Mobility Modeling. Given the signal S u (t) and the valid interpretation of each element, construct the user's mobility model in terms of a graph G u (ROI, Tr), where ROI = {ROI 1 , ROI 2 ...ROI n } is the set of all the re...

show abstract

Section: The Parameter Cursementioning

confidence: 97%

Section: Mobility Modelingmentioning

confidence: 99%

Section: Mobility Modelingmentioning

confidence: 99%

Section: Qs =mentioning

confidence: 99%

See 2 more Smart Citations

Capstone: Mobility Modeling on Smartphones to Achieve Privacy by Design

Kulkarni

Moro

Chapuis

et al. 2018

2018 17th IEEE International Conference on Trust, Security and Privacy in Computing and Communications/ 12th IEEE International

Self Cite

View full text Add to dashboard Cite

show abstract

“…Those pieces of information can be further used to project its future location statistically [7][8][9]. On the one hand, this type of prediction can be useful for certain purpose [10], but for a concerned privacy user, it is not good news.…”

Section: Introductionmentioning

confidence: 99%

Enhanced Internet Mobility and Privacy Using Public Cloud

Zhang

Durrësi

Durresi

2017

Mobile Information Systems

View full text Add to dashboard Cite

Internet mobile users are concerned more and more about their privacy nowadays as both researches and real world incidents show that leaking of communication and location privacy can lead to serious consequence, and many research works have been done to anonymize individual user from aggregated location data. However, just the communication itself between the mobile users and their peers or website could collect considerable privacy of the mobile users, such as location history, to other parties. In this paper, we investigated the potential privacy risk of mobile Internet users and proposed a scalable system built on top of public cloud services that can hide mobile user's network location and traffic from communication peers. This system creates a dynamic distributed proxy network for each mobile user to minimize performance overhead and operation cost.

show abstract

On the Inability of Markov Models to Capture Criticality in Human Mobility

Kulkarni¹,

Mahalunkar

Garbinato³

et al. 2019

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

We examine the non-Markovian nature of human mobility by exposing the inability of Markov models to capture criticality in human mobility. In particular, the assumed Markovian nature of mobility was used to establish a theoretical upper bound on the predictability of human mobility (expressed as a minimum error probability limit), based on temporally correlated entropy. Since its inception, this bound has been widely used and empirically validated using Markov chains. We show that recurrent-neural architectures can achieve significantly higher predictability, surpassing this widely used upper bound. In order to explain this anomaly, we shed light on several underlying assumptions in previous research works that has resulted in this bias. By evaluating the mobility predictability on real-world datasets, we show that human mobility exhibits scale-invariant long-range correlations, bearing similarity to a power-law decay. This is in contrast to the initial assumption that human mobility follows an exponential decay. This assumption of exponential decay coupled with Lempel-Ziv compression in computing Fano's inequality has led to an inaccurate estimation of the predictability upper bound. We show that this approach inflates the entropy, consequently lowering the upper bound on human mobility predictability. We finally highlight that this approach tends to overlook long-range correlations in human mobility. This explains why recurrent-neural architectures that are designed to handle long-range structural correlations surpass the previously computed upper bound on mobility predictability.

show abstract

A mobility prediction system leveraging realtime location data streams

Cited by 10 publications

References 3 publications

Capstone: Mobility Modeling on Smartphones to Achieve Privacy by Design

Capstone: Mobility Modeling on Smartphones to Achieve Privacy by Design

Enhanced Internet Mobility and Privacy Using Public Cloud

On the Inability of Markov Models to Capture Criticality in Human Mobility

Contact Info

Product

Resources

About