A Mobility Forecasting Framework with Vertical Federated Learning

The developments of Information and Communication Technology (ICT) and Internet of Things (IoT) are being used to enhance quality, performance and interactivity of urban services. The developments of Information and Communication Technology (ICT) and Internet of Things (IoT) are being used to enhance quality, performance and interactivity of urban services. Benefited from the widespread adoption of mobile devices, we can collect amount of mobile data for user mobility analysis. Mining hidden information from users' mobile data is important for builders of smart city to provide better location-based service. This paper focuses on two classical domain-independent prediction models and one improved Markov model that are capable of estimating the next location. By using 27-day-long traffic data of mobile network, we extract trajectories of 4914 individuals for experiments. We find that the original Markov algorithm has a better performance in resource consumption than LZ family algorithms, but its prediction accuracy is lower than prediction accuracy of LeZi Update and Active LeZi algorithm. In order to improve the prediction accuracy of Markov and overcome drawbacks of traditional prediction algorithms, we present a new method based on Markov, which considers both temporal and spatial factors. Extensive experiments demonstrate our improved method has a better performance in location prediction. In addition, we further study the relationship between prediction accuracy and trajectory's regularity, to identify the most suitable prediction algorithm for a trajectory.

In large-scale wireless sensor networks the data- gathering mechanism to introducing mobility into the network. . We consider the location service in a WSN, where each sensor needs to maintain its location information by 1) frequently updating its location information within its neighboring set of polling points in the network. In the previous work a mobile data collector, for convenience called an M-collector for gathering information from the sensor and send to sink node. M- collector could be a mobile robot or a vehicle equipped with a powerful transceiver and battery, working like a mobile base station and gathering data while moving through the field. An M-collector starts the data-gathering tour periodically from the static data sink, polls each sensor while traversing its transmission range, then directly collects data from the sensor in single-hop communications, and finally transports the data to the static sink. Since data packets are directly gathered without relays and collisions, the lifetime of sensors is expected to be prolonged. In this paper, we mainly focus on the problem of minimizing the length of each data-gathering tour and remove the redundancy for data gathering. We develop a stochastic sequential decision framework to analyze this problem. Under a Markovian mobility model, the location update decision problem is modeled as a Markov Decision Process (MDP). For the work with strict distance/ time and energy constraints, we consider utilizing single and multiple M-collectors and propose a data- gathering algorithm where multiple M-collectors traverse through several shorter subtours concurrently to satisfy the distance/time and energy constraints. To identify the redundancy of data from the sensor using Pearson auto correlated technique. Our proposed system for mobile data gathering scheme can improve the scalability and balance the energy consumption among sensors.

The combination of wireless mobile service and spatial information service introduces location-based services (LBS). LBS utilizes GIS, spatial positioning and embedded technologies in mobile computing environment to provide mobile terminal users location relevant information. The purpose of this paper is to design a framework supporting advanced LBS over 4G mobile-IP networks, with the transition of existing LBS services to an all mobile-IP based architecture. The architecture, compositions and challenges for LBS in 4G are discussed. The detailed functions of each component are presented. On the base of this, the research points in implementation including positioning methods, new service management model, mobile spatial data compression and transmission, data cache in mobile terminals are also proposed.

In the past few years, infectious diseases like COVID-19 have caused serious distress to the global society and the economy. To prevent its spread, the early detection and assessment of infectious diseases based on molecular tests or antigen testing of bodily have led to countless labor and material costs. Fortunately, with the rapid development of mobile localization and web techniques, the collected massive mobile trajectory data provides a promising solution for detecting positive cases. However, existing mobility data-driven infection case detection methods are limited in terms of modeling the complicated epidemic spreading processes and preserving user privacy of the mobility data. In this paper, we propose a novel graph convolutional networks (GCN) model for detecting high risk infection cases, where we incorporate a spatio-temporal hypergraph to model the complex interaction of individuals. Then, we elaborately design a privacy-preserving framework tightly coupled with the structure of the spatio-temporal hypergraph, which includes a mobility data obfuscation module to protect privacy and an accompanying confidence-aware mechanism to mitigate the consequent performance decline. Moreover, we introduce a causal propagation mechanism to further guarantee the temporal dependency and causal effect of the feature propagation in our spatio-temporal hypergraph, which introduces both the causal transform of node features and the causal gathering of edge features. Finally, extensive experiments on a large mobility dataset collected from location-based services (LBS) show that the proposed model improves the performance of infection case detection by at least \(12.47\%\) when compared with several widely adopted baselines. Besides, our code and datasets are available at the link 1 .

This special issue of the European Journal of Transport and Infrastructure Research (EJTIR) comprises three papers addressing some key issues in transport geographical research that use location tracking data. Preliminary versions of those papers were presented at the Mobile Ghent 2013 Conference. This event is part of a series of biannually held conferences on location-based services (LBS) that initially started in Tartu, Estonia. This editorial briefly points to the growing academic relevance of mobile phone tracking data and location-based services, describes the history and aim of the Mobile Ghent conference and gives a preview of the selected contributions.

Developments in Web and Wireless technologies have enabled the diffusion of large volumes of geospatial mobility data, and new challenges and opportunities have emerged for the GIScience research community, interested in extracting knowledge from these data. In most data analytics scenarios, well-known analytics platforms, such as KNIME or RapidMiner, offer practical general-purpose tools to data analysts. However, when dealing with mobility data, these platforms provide only limited support to some peculiar geospatial data manipulation tasks, thus forcing researchers and practitioners to manually implement significant portions of their pipelines, hindering productivity and replicability of the results. This paper presents a solution we are currently working on to support mobility data analysis. Our prototype, which we called KNOT (KNime mObility Toolkit), extends the KNIME Analytics Platform with a collection of new components specifically designed to support processing steps typical of mobility data, including map-matching, trajectory partitioning, and road network coverage analysis. To show the effectiveness of these components, we report also on how we applied them to perform a realistic analytical task on a real-world massive mobility dataset.

In pervasive computing, location-based services (LBSs) are valuable for mobile clients based on their current locations. LBSs use spatial window queries to enable useful applications for mobile clients. Based on skewed access patterns of mobile clients, non-flat wireless broadcast has been shown to efficiently disseminate spatial objects to mobile clients. In this paper, we consider a scenario in which spatial objects are broadcast to mobile clients over a wireless channel in a non-flat broadcast manner to process window queries. For such a scenario, we propose an efficient spatial air index method to handle window query access in non-flat wireless broadcast environments. The concept of largest empty rectangles is used to avoid unnecessary examination of the broadcast content, thus reducing the processing time for window queries. Simulation results show that the proposed spatial air index method outperforms the existing methods under various settings.

This tutorial presents the definition, the models and the techniques of location privacy from the data privacy perspective. By reviewing and revising the state of art research in data privacy area, the presenter describes the essential concepts, the alternative models, and the suite of techniques for providing location privacy in mobile and ubiquitous data management systems. The tutorial consists of two main components. First, we will introduce location privacy threats and give an overview of the state of art research in data privacy and analyze the applicability of the existing data privacy techniques to location privacy problems. Second, we will present the various location privacy models and techniques effective in either the privacy policy based framework or the location anonymization based framework. The discussion will address a number of important issues in both data privacy and location privacy research, including the location utility and location privacy trade-offs, the need for a careful combination of policy- based location privacy mechanisms and location anonymization based privacy schemes, as well as the set of safeguards for secure transmission, use and storage of location information, reducing the risks of unauthorized disclosure of location information. The tutorial is designed to be self-contained, and gives the essential background for anyone interested in learning about the concept and models of location privacy, and the principles and techniques for design and development of a secure and customizable architecture for privacy-preserving mobile data management in mobile and pervasive information systems. This tutorial is accessible to data management administrators, mobile location based service developers, and graduate students and researchers who are interested in data management in mobile information systems, pervasive computing, and data privacy.

In recent years, location based services (LBS) have become very popular. The performance of LBS depends on number of factors including how well the places are described. Though LBS enable users to tag places, users rarely do so. On the other hand, users express their interests via online social networks. The common interests of a group of people that has visited a particular place can potentially provide further description for that place. In this work we present an approach that automatically assigns tags to places, based on interest profiles and visits or check-ins of users at places. We have evaluated our approach with real world datasets from popular social network services against a set of manually assigned tags. Experimental results show that we are able to derive meaningful tags for different places and that sets of tags assigned to places are expected to stabilise as more unique users visit places.Keywordsplace taggingrecommendation systemsdata miningonline social networkslocation based services

Much research has recently been devoted to the data management foundations of location-based mobile services. In one important scenario, the service users are constrained to a transportation network. As a result, query processing in spatial road networks is of interest. We propose a versatile approach to k nearest neighbor computation in spatial networks, termed the Islands approach. By offering flexible yet simple means of balancing re-computation and pre-computation, this approach is able to manage the trade-off between query and update performance. The result is a single, efficient, and versatile approach to k nearest neighbor computation that obviates the need for using several k nearest neighbor approaches for supporting a single service scenario. The experimental comparison with the existing techniques uses real-world road network data and considers both I/O and CPU performance, for both queries and updates.

A key technical issue for human intention prediction from observed human actions is how to discover and utilize the spatio-temporal patterns behind those actions. Inspired by the well-known two-stream architecture for action recognition, this paper proposes an approach for human intention prediction based on a two-stream architecture using RGB images and optical flow. Firstly, the action-start frame of each trial of a human action is determined by calculating the L2 distance of the positions of the human joints between frames of the skeleton data. Secondly, a spatial network and a temporal network are trained separately to predict human intentions using RGB images and optical flow, respectively. Both early concatenation fusion methods in the spatial network and sampling methods in the temporal network are optimized based on experiments. Finally, average fusion is used to fuse the prediction results from the spatial network and the temporal network. To verify the effectiveness of the proposed approach, a new dataset of human intentions behind human actions is introduced. This dataset contains RGB images, RGB-D images, and skeleton data of human actions of pitching a ball. Experiments show that the proposed approach can predict human intentions behind human actions with a prediction accuracy of 74% on the proposed dataset. The proposed approach is further evaluated on the Intention from Motion (IfM) dataset, a dataset of human intentions behind human actions of grasping a bottle. The proposed approach achieves a prediction accuracy of 77% on the IfM dataset. The proposed approach is effective in predicting human intentions behind human actions in different applications.

Grid computing is an emerging technology providing the possibility to aggregate resources for the solution of computation- or data-intensive scientific tasks. Taking the evolution of mobile computing into consideration, new Grid concepts are conceivable, fully exploiting the advantage of mobile devices and ubiquitous access. By decoupling resource availability from the core grid infrastructure and hardware, the user has always the same computational power, data or storage available, regardless of a device or location. Thus restricted capabilities of thin clients can be extended and new fields of application can be made accessible. The key concept is “The invisible grid” – the grid environment should just be there for the use of applications in science, business, health care, environment, or culture domains. Having this concept in mind, the following scenario is conceivable: Equipped with your mobile phone, which you always have with you, you are walking around and are taking a picture of an object you are interested in. You are sending the picture to the grid, where the visual information is extracted. After the analysis, information about the captured object is sent to you. Thus you have a search engine on a visual base at your permanent disposal, information captured as seen by your eyes – without the need of textural translations or the need to know the object’s name or ID in order to retrieve information about it. Realizing the scenario above, the user obtains a smart tool, easing information retrieval considerably by making use of ubiquity in combination with grid computing. But the scenario has even more potential in terms of pervasiveness. The use of mobile devices can provide a user with additional location bound information. With a portable device the user is able to access location-based services or to collect environmental information to be processed within a grid. At this stage research activities in the field of pervasive computing comeinto play. Pervasive computing pursues the goal to enhance the environment with sensors and smart objects in order to provide the user with suitable context-based and/or location-based services. Expanding the introduced setting with the capabilities from pervasive computing, the following scenario is conceivable: You are an invited speaker on a conference and you are moving through the rooms of the venue. All rooms are equipped with cameras covering all perspectives of view. You are looking at a person from whom you want to know the research interests. You flick with your finger, to capture the camera picture from your perspective. The picture is processed within the grid and the ambient display next to you shows the requested information.

Recently, ubiquitous computing / networks have been studied actively. These networks provide services depending on real environments of mobile nodes. Especially, we expect location-based services (LBSs), which rely on location of mobile nodes, are anticipated to come into wide use in the future. High-value LBSs require cryptography to ensure security. Here, cryptographic LBSs comprise a key management function (e.g. key sharing with nodes) and a location management function (e.g. location verification of nodes). Cooperation between key and location management functions realizes cryptographic LBSs. However, these functions have mostly been studied individually. This study indicates that cryptographic LBSs are insecure if the cooperation is incomplete, and proposes a method of constructing secure cryptographic LBSs.KeywordsMobile NodeLocation ManagementLocation EvidenceManagement PathClient AuthenticationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

This paper presents HERMES, a prototype DB engine that defines a powerful query language for trajectory databases, which enables the support of mobility-centric applications, such as Location-Based Services (LBS). HERMES extends the data definition and manipulation language of Object-Relational DBMS (ORDBMS) with spatio-temporal semantics and functionality based on advanced spatio-temporal indexing and query processing techniques. Its implementation over two ORDBMS and its utilization in various domains proves the expressive power and applicability of HERMES in different application domains where knowledge regarding mobility data is essential. As a proof-of-concept, in this paper HERMES is applied to a case study related with vehicle traffic analysis, demonstrating its flexibility and usefulness for delivering custom-defined LBS.

Nearest-neighbor search, which returns the nearest neighbor of a query point in a set of points, is an important and widely studied problem in many fields, and it has wide range of applications. In many of them, such as sensor databases, location-based services, face recognition, and mobile data, the location of data is imprecise. We therefore study nearest-neighbor queries in a probabilistic framework in which the location of each input point is specified as a probability distribution function. We present efficient algorithms for - computing all points that are nearest neighbors of a query point with nonzero probability; and - estimating the probability of a point being the nearest neighbor of a query point, either exactly or within a specified additive error.