Michela Papandrea

Location prediction and mobility modelling for enhanced localization solution

Current and future mobile applications massively exploit the knowledge of the user’s location to improve the offered services. However, user localization is by far one of the oldest and most difficult issues, due to its dynamism and to unavailability of some technologies in indoor environments. The enhanced localization solution (ELS) proposed in this paper is an innovative self adaptive solution that smartly combines standard location tracking techniques (e.g., GPS, GSM and WiFi localization), newly built-in technologies, as well as human mobility modelling and machine learning techniques. The main purposes of this solution are: to reduce the impact the service has, on the mobile device resources usage (mainly the battery consumption), when it is asked to provide a continuous localization; to help in preserving the privacy of the user, by running the whole system on the mobile device, without relying on a back-end server; and furthermore, to offer an ubiquitous coverage. The aspects mainly explored in this paper are: location prediction and mobility modelling, required to optimally estimate the current location with ELS. We are finding that people tend to move, for most of the time, among a limited set of places and that this can be modelled with a user prediction graph, which is further used to predict the next movement. Performing experiments on real users data, we show that the proposed prediction and the mobility model method of ELS are able to successfully predict the next location, even if we do not account for time features.

How many places do you visit a day

People mobility enormously augmented in the last decades. However, despite the increased possibilities of fast reaching far places, the places that a person commonly visits remain limited in number. The number of visited places of each person is regulated by some laws that are statistically similar among individuals. In our previous work, we firstly argued that a person visit most frequently always few places, and we confirmed that by some initial experiments. Here, in addition to further validating this result, we build a more sophisticate view of the places visited by the people. Namely, on top of our previous work, which identifies the class of Mostly Visited Points of Interest, we define two next classes: the Occasionally and the Exceptionally Visited Points of Interest classes. We argue and validate on real data, that also the occasional places are very limited in number, while the exceptional ones can grow at will, and by the analysis of the classes of visited points we can distinguish the type of users mobility. This paper firstly demonstrates this property in large experimental scenario, and put the basis for new understanding of people places in several areas as localization, social interactions and human mobility modelling.

On the properties of human mobility

Visited locations are classified in 3 main categories according to their relevance.People visit regularly just few places where they spend most of their time.People commute between places based on their time (not spatial) distance.HOME and WORK places are in the set of few places mostly visited.Mostly visited places semantic inference is based on user mobility/behavior analysis. The current age of increased people mobility calls for a better understanding of how people move: how many places does an individual commonly visit, what are the semantics of these places, and how do people get from one place to another. We show that the number of places visited by each person (Points of Interest - PoIs) is regulated by some properties that are statistically similar among individuals. Subsequently, we present a PoIs classification in terms of their relevance on a per-user basis. In addition to the PoIs relevance, we also investigate the variables that describe the travel rules among PoIs in particular, the spatial and temporal distance. As regards the latter, existing works on mobility are mainly based on spatial distance. Here we argue, rather, that for human mobility the temporal distance and the PoIs relevance are the major driving factors. Moreover, we study the semantic of PoIs. This is useful for deriving statistics on peoples habits without breaking their privacy. With the support of different datasets, our paper provides an in-depth analysis of PoIs distribution and semantics; it also shows that our results hold independently of the nature of the dataset in use. We illustrate that our approach is able to effectively extract a rich set of features describing human mobility and we argue that this can be seminal to novel mobility research.

Enhanced Localization Solution

The high potential of new generation smartphones in terms of resources capability as well as availability of embedded sensors and radio interfaces has opened new perspectives to developers, which came out with a massive number of applications for mobile phones. Most of them are location based, where location data is used either as main or as auxiliary information. However, the location service is power hungry. Thus, a major challenge is the optimization of the trade-off between resources usage of the location service and its accuracy. We introduce the architecture of the Enhanced Localization Solution (ELS), an efficient localization strategy for smartphones which smartly combines the standard location tracking techniques (e.g., GPS, GSM and WiFi localization), the newly built-in technologies, as well as Human Mobility Modelling and Machine Learning techniques. This solution aims to provide a continuous and ubiquitous service while reducing the impact on the devices resources usage.

A lightweight localization architecture and application for opportunistic networks

Mobile communication is increasingly becoming part of everyday life. Haggle (www.haggleproject.org) is an innovative project that exploits human mobility for creating an invisible social network to enable communication in the presence of intermittent network connectivity. Mobile devices equipped with network interfaces, may in fact (temporarily) be out of range of the Internet infrastructure, and thereby unable to access Internet services. However, these devices may often be in range of other networked devices, or pass by an area where some type of connectivity is available (e.g. WiFi hotspots) and may be able to perform useful networking transactions involving local exchanges of data, exploiting, in this way, the opportunistic communication. One of the mechanisms for realizing the opportunistic communication is by taking advantage of context information (such as location) for correlating on-line social and physical networks. This, together with the need for information regarding devices locations and the ‘pervasive’ presence of wireless networks in indoor environments, suggested us to implement a lightweight application able to localize mobile devices, exploiting opportunistic networking.

A smartphone-based energy efficient and intelligent multi-technology system for localization and movement prediction

Nowadays we are assisting to a noticeable proliferation of new generation smart-phones, as well as to the growth of mobile applications development. We are fairly surrounded by a huge number of proactive applications, which automatically provide users with relevant information exactly at the right time and at the right place when they need it. This ability is generally given by the exploitation of context information, and mainly Location Information. However, a location service running onto a smartphone has to deal with a great challenge, which is to manage the trade-off between the services resources usage and its accuracy. In my work I investigate efficient localization strategies: these include, in addition to the standard location tracking techniques, the support of other technologies already available on mobile phones (i.e., sensors), as well as the integration of either Human Mobility Modelling and Machine Learning techniques. The main purposes of this work are: to reduce the impact that the service has on the devices resources usage in the case of continuous localization; to preserve the privacy of the user by running the whole system on the mobile device without relying on a back-end server; and to offer an ubiquitous coverage.

Multimodal ubiquitous localization: a GPS/WiFi/GSM-based lightweight solution

Localization is a key support service for all users of mobile devices. We envision a lightweight localization service for Smartphones that overcomes the limitations of existing GPS-based approaches, namely their high energy cost and their issues with coverage. Our localization service uses GPS only if available and necessary, and employs other methods, such as WiFi and GSM, to achieve its goal of ubiquitous and energy-efficient operation. In this paper, we discuss our preliminary results as well as our future research directions.

On the key features in human mobility: Relevance, time and distance

Our era of increased people mobility requires a better understanding of how people move, how many are the places that a person commonly visits, how people move among them. In our previous work we have shown that the number of visited points of interest (PoIs) of each person is regulated by some laws that are statistically similar among individuals, and we gave some classification of them in terms of their relevance. We investigate here the variables that characterize the way humans move among PoIs, and in particular the spatial and temporal distances between PoIs and the PoIs relevance. With this respect, most of the existing work on mobility, especially on its modeling, is based only on spatial distance, while we argue here that for human mobility the temporal distance and the PoIs relevance are essential factors. Also, by comparing the geographical and temporal distances between consecutive PoIs, we observe a smoother and more linear trend in the temporal case. The results suggest that the time, rather than physical distance, could represent a better measure of distance among PoIs, mainly due to the transportation facilities. We present here our analysis of this aspect, considering two different datasets, to show the validity of our results independently of the nature of the dataset under consideration, and to extract considerations on different scenario scales. We argue that these results can be seminal to novel mobility research.

Synergistic user ↔ context analytics

Various flavours of a new research field on (socio − )physicalorpersonalanalytics have emerged, with the goal of deriving semanticallyrich insights from people’s low-level physical sensing combined with their (online) social interactions. In this paper, we argue for more comprehensive data sources, including environmental and application-specific data, to better capture the interactions between users and their context, in addition to those among users. We provide some example use cases and present our ongoing work towards a synergistic analytics platform: a testbed based on mobile crowdsensing and IoT, a data model for representing the different sources of data and their connections, and a prediction engine for analyzing the data and producing insights.

VIVO: A secure, privacy-preserving, and real-time crowd-sensing framework for the Internet of Things

Smartphones are a key enabling technology in the Internet of Things (IoT) for gathering crowd-sensed data. However, collecting crowd-sensed data for research is not simple. Issues related to device heterogeneity, security, and privacy have prevented the rise of crowd-sensing platforms for scientific data collection. For this reason, we implemented VIVO, an open framework for gathering crowd-sensed Big Data for IoT services, where security and privacy are managed within the framework. VIVO introduces the enrolled crowd-sensing model, which allows the deployment of multiple simultaneous experiments on the mobile phones of volunteers. The collected data can be accessed both at the end of the experiment, as in traditional testbeds, as well as in real-time, as required by many Big Data applications. We present here the VIVO architecture, highlighting its advantages over existing solutions, and four relevant real-world applications running on top of VIVO