SNET Projects

Context Data Cloud

Mar 2013 – Feb 2015

Partners Technische Universität Berlin, Telekom Innovation Laboratories
Funding by Software Campus, Federal Ministry of Education and Research (BMBF)

The Context Data Cloud has been chosen as one of the most innovative IT projects in 2013 to be funded within the Software Campus initiative.

It is a service enabler platform (run by a mobile network operator) that interconnects context providers, consumers as well as context-aware service developers for providing semantically enriched context-aware services. It is based on Linked Data and enables developers of context-aware services to easily realize innovative applications by making use of a set of novel interfaces (provided by the network operator).

Context Data Cloud

This platform mainly consists of four building blocks: The OpenMobileNetwork provides approximated and semantically enriched mobile network and WiFi access point topology data enabling the development of sophisticated context-aware services. Linked Crowdsourced Data, on the other hand, enriches static location entity (e.g., POIs or events) information by dynamic contextual parameters for realizing various location analytics scenarios. The Context Management is responsible for context search and enrichment utilizing the LOD Cloud, whereas the Context Consumer Interfaces provide developers the possibility to implement services on top of the platform.

Context Data Cloud for Android (CDCApp)

The Context Data Cloud for Android (CDCApp) is a location-based community app offering a set of semantic services to users such as a Friend Tracker or a Popular Places Finder. Users can check-in to certain points of interest in their vicinity providing information about their favorite or frequently visited locations to other users within their social community. These services are based on the Context Data Cloud platform and exploit the rich set of data available within the LOD Cloud.

Please do not hesitate to download the app at the Google Play Store!

Context Data Cloud for Android (CDCApp)

This app collects personal location preference data (e.g., favorite or frequently visited locations) in combination with the current contextual situation (e.g., the weather when checking in to a point of interest). A totally anonymized and aggregated version of this dataset is published in Linked Data format as Linked Crowdsourced Data and is interlinked to other datasets within the LOD Cloud.

Linked Crowdsourced Data

Geospatial data that is published within the LOD Cloud (e.g., LinkedGeoData) mostly consists of static information, such as a name, geo coordinates, an address, or opening hours. Enriching this data with location-specific crowdsourced information (e.g., checkins, ratings, or comments) as well as the contextual situation in which users visited a place (e.g., time of the day, weather condition, dwell time, or holiday) and interlinking this information with other datasets will enable location analytics scenarios within the LOD Cloud.

Linked Crowdsourced Data is - as the name implies - a crowdsourced dataset for real user location preferences. People using the CDCApp contribute to this dataset by checking into different locations. This information is accompanied by contextual information, such as the weather condition during the check-in, the WiFi access points that were measured in this area or whether the check-in occured during a regional holiday.

This dataset is published as Linked Data in RDF format through our SPARQL endpoint. An aggregated and anonymized visualization of the data is provided on our Location Analytics Map.

Location Analytics Map

Semantic Positioning Solutions

The location of a user is the fundamental factor shaping location-based services (LBS) and is usually computed solely in terms of WGS84 coordinates without taking the semantics of the location into consideration. Furthermore, traditional proactive LBS require manual effort in defining geofences beforehand around all relevant location entities, which is very time-consuming and not feasible in the LBS development process. Another challenge is related to Geocoding Services where incomplete or ambiguous address data often leads to wrong results.

In order to overcome these drawbacks, we provide several Semantic Positioning Solutions as integral parts of the Context Data Cloud platform that overcome the limitations of classic geocoding as well as geofencing methods and add semantic features to proactive self-referencing and cross-referencing LBS. The applicability of our Semantic Tracking and Semantic Geocoding approaches is showcased within the Friend Tracker as well as Popular Places services of the CDCApp and our OpenMobileNetwork Geocoding Web Interface.


Dec 2011 – Present

Developer Abdulbaki Uzun

The OpenMobileNetwork is an open platform that provides approximated and semantically enriched mobile network and WiFi access point topology data published as Linked Data in RDF format.

OpenMobileNetwork - Map

Since 2012, users from all over the world contribute to this dataset by running one of our dedicated Measurement Clients on their mobile devices. This crowdsourced data is used to approximate mobile network as well as WiFi access point topologies worldwide. The processed data is structured according to a comprehensive Ontology in order to provide information about base station and WiFi access point locations, their coverage areas, neighboring cell relations and dynamic data, such as traffic, service usage, and number of users. It is made available to the public through a SPARQL endpoint and a meta description is provided using the Vocabulary of Interlinked Datasets (VoID).

In compliance with the principles of Linked Data, many interlinks to other LOD Cloud datasets such as LinkedGeoData or DBpedia enable various applications ranging from Semantic Location-based Services to a sophisticated Power Management in Mobile Networks.

OpenMobileNetwork for Android (OMNApp)

OpenMobileNetwork for Android (OMNApp) is a background service that actively collects mobile network data in a crowdsourcing approach in order to estimate and derive base station and WiFi hotspot positions. It further acquires live measurement data, such as the total traffic produced on smartphones or service usage information, in order to enable the modeling and visualization of the current/historic state of the network.


Jewel Chaser

Jewel Chaser is a location-based game in which you collect virtual jewels by moving to locations in the real world. While you play the game, mobile network and wireless access point data is collected for the OpenMobileNetwork in order to enrich the existing database.

Jewel Chaser   Jewel Chaser

In order to play it, just click on Start Collecting, check the map, and go towards one of the symbols shown on the map. As soon as you are close enough to a jewel, it will be collected automatically. However, you need to have the map open in order to collect the jewel. This means you can check where the jewels are, close the app, and open it later again as soon as you are close to one of the jewels. In order to collect jewels, your GPS has to be turned on.

Communicate Green

Jan 2011 – Apr 2014

Partners Technische Universität Berlin, Telekom Innovation Laboratories, Ericsson GmbH, Fraunhofer Gesellschaft zur Förderung der angewandten Forschung e.V., Universität Paderborn
Funding by Federal Ministry for Economic Affairs and Energy (BMWi)

In order to fulfill today’s high demands on mobile network usage, mobile network providers in Germany have around 123.000 base stations working 24 hours a day, 7 days a week 7 with a total power consumption of approx. 1455 GWh per year. The permanent availabilty of those network components causes a significant energy consumption. Through an adaptive, context-aware and technology-comprehensive power management in mobile networks, a considerable amount of energy can be saved by maintaining the high quality of experience at the same time. There are a lot of contextual information present in mobile network components and end devices, which can help to calculate decisions for a dynamic de- and reactivation of network components (e.g., base stations).

The main focus of our work in this project is the development of a context entity that aggregates and processes contextual information in order to calculate decisions for network optimizations. By applying various context acquisition approaches, both the network as well as the user view will be included into the power management process.

Communicate Green

ComGreen Showtable

ComGreen Showtable

For the purpose of the project, Telekom Innovation Laboratories and Service-centric Networking developed a test environment consisting of several components in order to visualize the effects of energy management algorithms for mobile networks under realistic conditions. This test environment comprises a multi radio access technology (Multi-RAT) testbed, which is a professional testbed suite consisting of a number of different radio access technology nodes. The second component is a Showtable that demonstrates an energy management scenario based on contextual information gathered from network components as well as mobile devices. These contextual parameters are collected, evaluated and delivered to the algorithms by the Context Manager.

The scenario demonstrated by the Showtable uses WiFi access points and focuses on the de- and reactivation of unused network components and the intelligent handover of the active users. As an exemplary application, video streams are used to visualize the status of the network on an external screen and to display real-time information about traffic load and power consumption. The deployed algorithm is a threshold-based policy, which switches off one of the access points in case that aggregated traffic load of both access points is below a defined threshold. At the same time, the energy management algorithms guarantee a consistently high quality of experience perceived by the user.

OpenMobileNetwork for ComGreen Demo

The OpenMobileNetwork for ComGreen Demo visualizes how interlinked semantic data can be useful in order to establish a power management in mobile networks. For this purpose, we extended the OpenMobileNetwork Map by several power management scenarios. The ComGreen checkbox visualizes the capability of choosing alternative mobile network cells for users to be handed over in order to shutdown candidate cells, whereas the Service Usage checkbox determines cells with a characteristic service usage profile.

OpenMobileNetwork for ComGreen Demo

Please feel free to have a look at the special OMN for ComGreen Demo. You can see live traffic produced by your smartphone, if you run OpenMobileNetwork for Android (OMNApp), which is available for download at Google Play Store.

OpenMobileNetwork Predictions Visualizer

In order to implement energy-efficient reconfiguration mechanisms in mobile telephony networks, it is essential to anticipate traffic and user hotspots, so that a network's configuration can be adjusted in time accordingly. The mechanisms that we have employed can estimate the future amount of traffic as well as user movements on a network cell level based on data from the OpenMobileNetwork.

OpenMobileNetwork Predictions Visualizer

The OpenMobileNetwork Predictions Visualizer showcases several prediction scenarios including Traffic Predictions, Individual Movement Predictions, and Group Predictions.