Philipp Svoboda is a senior scientist working at TU Wien, with a research focus on the performance aspects of mobile cellular technologies.
He is currently examining the feasibility of using crowdsourcing to conduct performance measurements on 4G and 5G mobile networks.
His research aims to establish a common framework for evaluating the performance of mobile networks, guaranteeing reliable and fair connectivity for end-users.
Key fingerprint = C850 2F3F E47C 2B7F 295C 17EB 92EF 91EC FAC7 863B
In 2015 he has been elevated to be an IEEE Senior member.
Research-wise he has started research work with the COMET center FTW Vienna focusing on the user behavior in the mobile core networks of various operators in Austria.
Today he is involved several national projects (FFG, …) and was leading work-packages in several european FP7/H2020 projects (LOLA, ADWICE, …). The current focus is on modelling M2M communication systems for 4G and 5G networks.
In his spare time he is investigating traffic models for online games and measuring service performance in 5G networks. His current research interests include simple traffic generation, statistical analysis of IP level information and modeling of new services, specifically on all kinds of mobile networks.
- OEBB TS: Improvement of service quality for mobile cellular networks on board of railroad cabins
- A1 Telekom Austria AG: Convergence of mobile and fixed wired access technologies
- FFG, Mc. HypaMiner: Crowdsourcing performance measurements for mobile cellular users
- H2020, Open Call: FaliCap / MONROE
- A1 Telekom Austria AG: Performance optimizations in future home networks
Past Project Activity:
The project Advanced Wireless Technologies for Clever Engineering (ADWICE) is aimed to create a strong partnership between the research center of Sensor, Information and Communication Systems (SIX, Czech Republic) and Vienna University of Technology (TUW, Austria). QS World University Rankings sets TUW on the 91st position among engineering faculties worldwide. The partnership will result in the transfer of excellence in research from TUW to SIX.
SIX is located in the region of South Moravia. The regional innovation strategy 2014-2020 (RIS) identifies (1) electrical engineering, (2) information technologies, (3) mechanical engineering and (4) life sciences as dominant sectors of the regional economy. SIX contributes (1) to (3).
Smart specializations of South Moravia identified by RIS are (1) Advanced manufacturing & engineering, (2) Accurate instruments, (3) Hardware & software, (4) Pharmaceuticals, medical care & diagnostics, and (5) Aeronautical technologies. Wireless technologies can find exploitation in all these specializations which is documented by letters of intent provided by companies.
The ADWICE project consists of work-packages covering (1) Sensor systems, (2) Signal processing, (3) Radiofrequency applications, (4) Mobile communications and (5) Cyber security. Work-packages:
– Are co-supervised by a TUW leader and a SIX one;
– Contribute to (1) Smart cities, (2) Mobility for growth, and (3) Digital security;
– Are associated with companies.
The ADWICE project will result in a sustainable network comprising companies, SIX and TUW. The network will strengthen the innovation potential of companies thanks to the applied research of SIX and TUW. The network will cooperate on common research, education and dissemination. Operation of the network will be financed from private sources (contributions of companies) and public ones (national funds, HORIZON 2020). Research in the initial phase (2015-2019) will be funded by the National Sustainability Program.
The focus of LOLA is on access-layer technologies targeting low-latency robust and spectrally-efficient transmission in a set of emerging application scenarios. We consider two basic types of wireless networks, namely long-range LTE-Advanced Cellular Networks and medium-range rapidly-deployable mesh networks. Research on low-latency transmission in cellular networks is focused firstly on transmission technologies in support of gaming services which will undoubtedly prove to be a strategic revenue area for operators in the years to come. Secondly, we also consider machine-to machine (M2M) applications in mobile environments using sensors connected to public infrastructure (in trains, busses, train stations, utility metering, etc.). M2M is an application area of extremely high growth potential in the context of future LTE-Advanced networks. A primary focus of the M2M research is to provide recommendations regarding PHY/MAC procedures in support of M2M to the 3GPP standardization process. The rapidly-deployable mesh topology component addresses M2M applications such as remote control and personnel/fleet tracking envisaged for future broadband civil protection networks. This work builds upon ongoing European research in this important area. Fundamental aspects of low-latency transmission are considered in addition to validation on real-time prototypes for s subset of the considered application scenarios. The cellular scenario validation is carried out using both live measurements from an HSPA test cell coupled with large-scale real-time emulation using the OpenAirInterface.org emulator for both high-performance gaming and M2M application. In addition, a validation testbed for low-layer (PHY/MAC) low latency procedures will be developed. The rapidly deployable wireless mesh scenario validation makes use of the real-time OpenAirInterface.org RF platform and the existing FP6 CHORIST demonstrator interconnected with commercial M2M equipment.
3G mobile cellular networks have become an integral part of everyday communication serving a wide range of heterogeneous applications. Business-critical services such as e.g., remote payment services, machine-to-machine communication or traffic fleet management applications may coexist in the network with leisure applications such as e.g., video streaming or social networking. Due to the growing importance and wide-spread usage of 3G networks, assuring the availability and robustness of such networks has become a task of critical importance. 3G networks are inherently complex systems and are subject to heterogeneity induced by the variety of terminal types and their different capabilities as well as by a diverse set of network equipment and access technologies. Moreover, the steady increase of transferred traffic volume and the continuous evolution towards higher access bandwidth capacities necessitate permanent updates of the underlying network infrastructure. This evolving nature and complexity expose 3G networks to problems and errors, making the online monitoring of network behaviour and the timely detection and reporting of network problems highly desirable for 3G network operators.
The DARWIN4 project is placed along the research path originating from the previous DARWIN projects and hence it will also rely on research results obtained during the previous projects (e.g. novel network performance signals or the large-scale research database, both having been developed in DARWIN3). The research of DARWIN4 aims at advancing the current state of the art by focusing on the following innovations:
- The emerging of “always-on” smart phones and the increased usage of mobile apps presumably result in an increased demand of signalling capacity. Based on a thorough investigation of signalling traffic patterns present in the network and on an analysis of how these impact the available signalling capacity, DARWIN4 will focus on the development of innovative methods for detecting congestion of network-wide signalling capacity and also other types of anomalies such as macroscopic signalling attacks or malfunction of network equipment.
- The availability of network-wide passive monitors allows for a multi-dimensional view on the health status of the 3G network. DARWIN4 will investigate novel online and near real-time alarming schemes for the detection of network faults such as equipment malfunctions or congestion of signalling plane and user plane capacity. The focus of this activity is to study the technical feasibility of near real-time alarming based on the availability of aggregate traffic statistics. Accordingly, the work on real-time alarming will be based on the availability of a large-scale, high-performance research database and it will also rely on passively recorded network-wide performance signals, both having been initiated in the previous DARWIN projects. The proposed alarming scheme will be assessed based on the operationally relevant criteria coverage, short detection delays and detection accuracy.
- The emerging use of M2M applications and their particular characteristics on the user plane as well as on the signalling plane necessitates measures to detect M2M terminals and to gain control on their impact on the available signalling traffic capacity. Accordingly, DARWIN4 will focus on the characterisation and detection M2M-related traffic including a study of its impact onto the available signalling capacity.
Recent literature has demonstrated that an inefficient usage of the road network causes non-negligible economic losses in developing and developed countries and affects air pollution and greenhouse gas emission. Sensing and providing real-time road traffic information would inherently optimize traffic flows and reduce congestion events. However, the costs of deploying new sensors covering the whole road network are unsustainable for both road operators and municipalities. The RoadCell project proposes the use of an operational and widely available technology such as GSM/UMTS as a valid low-cost alternative for sensing the road condition in large geographical areas, bringing economical, operational, and environmental benefits.
Currently offered Thesis
The following tasks are available at the TU Vienna / ICT in the group of mobile communications. They all are focused on mobile cellular network and network monitoring in either passive or active way. Even if not explicitly named you will most likely work with traces from a 4G network on either lower signaling layers up to the IP/application layer.
The ideal candidate would have fun and basic knowledge in working with TCP / UDP / network traces. Depending on the section the possible projects have either more practical or theoretical elements.
I have prepared an overview of current topics available for project works. This list is not complete, so feel free to browse – in case of interest please contact us (email@example.com).
Thesis/project works in this section contain substantial amount of basic hardware tinkering (soldering, wiring and stuff) in the first step of the thesis. In the following steps you will gather measurements results and learn how to collect and process traces. In the final step you will evaluate these traces and find models or optimzations.
Audio One-Way-Delay Measurement Setup and Model (Thesis Master)
The goal of this project is to build a hardware setup capable to measure audio delay in a cellular mobile network, example setup depicted in the attached figure below. After this initial step the audio delay shall be evaluated for different scenarios in the network, e.g. with different operators. The audio transmission in GSM and UMTS is based on an AMR coding scheme with packatization introducing additional distortion to the original signal the next step is the design of a test impulse allowing for high resolution. Finally the Voice Quality shall be evaluated on the system and a model for the one-way delay derived from the measurements.
Raspberry Pi (alike: TI BeagleBone Black) Measurement Node (Thesis Master)
The release of potent hardware in ultra slim format, e.g., RaspberryPi and BeagleBone Black, allow for new type of network measurement nodes. In this master thesis the goal is to develop such a system capable of running basic network performance measurements in a autarcik and mobile way, e.g., drive one-way delay from a USB 4G dongle. For this purpose it is needed to setup some hardware as a preliminary step, e.g., GPS connection with the board and install existing software. In the following the student shall develop/port own measurement tools for the network scenarios and derive according models and improvements. As the topic is new there are numerous sub tasks to select from in this thesis listed below:
· SubTask: Engery Measurement
· SubTask: Network Patern Performance
· SubTask: Network Pattern Delay
· SubTask: GPS TimeSync
· SubTask: Master/Slave Autonode
· SubTask: Audio Quality Measurement
· SubTask: QoS und MOS Mapper
· SubTask: Output on Display
· SubTask: Input from Keys
Network Simulator Software Tasks
Thesis/project works in this section contain work with network simulators. The projects will be mainly conducted at a PC in the office. In the first step you will learn about the specific simulator in use and setup the test scenario for your work. In the second step you will run and simulate these setups collecting several runs of results. In the final step you will analyze the results in Matlab and or other mathematical tools.
Network Simulator (OpenAir) (Thesis Master)
The OpenAir simulation environment allows the user to simulate a LTE network from PHY up to the IP layer. In this thesis the student will setup a simulation environment presenting a LTE network with hundreds of active M2M nodes. The relating traffic code is already part of the network simulator. The task for the student is to setup, run and evaluate the simulations for different scenarios, e.g., low and high mobility nodes. These results will be compared to measurements conducted in the real network. In the final stage a real world transmission based on the traffic patterns collect on the network interface shall be done in the MIMO testbed of the university.
· Simulate M2M Nodes
· Simulate Traffic on Interfaces – write module / results / optimization
· Show load on different interfaces for different scenarios
· Make a MIMO Testbed Transmission – provide input stream
Signal Processing in Mobile Communications
Thesis/project works in this section contain substantial signal processing tasks to do. You will mainly work with Matlab and a PC on a high level of network traces, e.g., IP traces form a router. The preliminary steps of this work contain the study of appropriate algorithms for the task in the work. The following step will be to implement the algorithms into Matlab and to analyze the results based on toy models/ artificial traces. If this step succeeds you will test the algorithms on real-traces and learn how to process recorded traces from real world environments.
Traffic Clustering and Grouping (Thesis Master)
· Test of different Algorithms for Clustering and Grouping
o t-Distributed Stochastic Neighbor Embedding
o Cluster M2M Traffic Patterns
· Test: KNIME for big-data in network monitoring
· Test: RLab
Channel Sounding in Packet Switched Networks (Thesis Master)
· Signal Data Processing for packet switched networks
· How do time series change through a network box with other traffic inside – what does this tell about the the other traffic?
Key fingerprint = C850 2F3F E47C 2B7F 295C 17EB 92EF 91EC FAC7 863B