Académique Documents
Professionnel Documents
Culture Documents
Stefan Heiss
Assignment of tasks
The goal of this Project is to connect a TPM with an I2C-Interface with a NIOS II development board. In particular, in carrying out this project work you should learn TPM basics, integrate an I2C IP-core in Alteras NIOS II system, develop a basic software module to test the communication with the TPM.
Prerequisites:
Contact:
Prof. Dr. Stefan Heiss stefan.heiss@hs-owl.de (Supervisor) M.Sc. Stefan Hausmann stefan.hausmann@hs-owl.de
Assignment of tasks
The goal of this project work is to evaluate layer 2 tunneling protocols. In particular, in carrying out this project work you should identify tunneling protocols which are suitable to be used with IPSec, build up a test system to evaluate the operation of these protocols, measure the performance overhead introduced by the application of tunneling protocols.
Prerequisites:
Contact:
Prof. Dr. Stefan Heiss stefan.heiss@hs-owl.de (Supervisor) M.Sc. Stefan Hausmann stefan.hausmann@hs-owl.de
Assignment of tasks
Identify and evaluate available test cases (NASL scripts) from an OpenVAS standard installation that are related to vulnerabilities described in VuTAT-DOC. These test cases should be generic in the sense that they do not depend on specific properties (operating system, etc.) of the components to be tested. Develop new NASL scripts to test vulnerabilities described in VuTAT-DOC, but not sufficiently covered by scripts contained in the OpenVAS standard installation. If necessary, integrate some functionality in the OpenVAS implementation that allows reacting on external events (failure status indicated by some external test device).
Prerequisites:
Contact:
Prof. Dr. Stefan Heiss stefan.heiss@hs-owl.de (Supervisor) Dipl.-Ing. Andreas Schmelter andreas.schmelter@hs-owl.de
Description
Spectral hole detection provides the base for cognitive decision making. Traditional hole detection methods are dumb, in a sense that they merely provide knowledge about the instant occupancy of the sensed band and amplitude of the primary signal. Consequently, underlying radio platform can not fully exploit orthogonal oppertunities available in coexisting environments. As an example, a radio system capable fo choosing CDMA dont need to switch frequency or find temporal gaps in current channel if it can know that a CDMA based primary user is active. It can simply choose an orthogonal code to safly coexist in an already occupied channel. Similarly, there can be many other orthogonal oppertunities with repect to polarization, diversity, MIMO etc. all of which needs enough situational knowledge. Intellegent hole detection provides this situational knowledge, which requires identification of coexiting systems and inclusion of a radio knowledge base (RKB) for documented information of primary systems of interest. There are two important constituents of IHD, one is situational knowledge and the other one is a radio platform capable to tune available orthogonal oppertunities. This project is proposed to implement a 2-FSK based system which can optionally choose CDMA and can tune its frequency and transmission power and implements traditional energy detection based energy sensing (but not signal classification). Eventually, a reconfiguerable platform will be availabe that would be able to exploit frequency, time, power and code dimensions. While doing experiments for comparison of dumb hole detection (DHD) and IHD, perticular coexisting scenerio will be created and hardcoded to the IHD based system as proxy to classification based situational knowledge. A USRP2 based system should be used for implementation.
Contact:
Area: Robotics / Communication Title: Energy consumption / billing monitor for charging mobile devices
Description
Mobile robots are typically powered electrically and need to be recharged frequently. This charging can also be realized without a wired connection. The task is to design an energy counter for such a wireless charging station, keeping track of the supplied robots and accumulating the individual energy consumption. This includes implementing a serial interface to the charger, reading specific parameters and visualizing the supplied energy on a PC.
Contact:
Contact:
Contact:
Tasks:
Help to evaluate and develop the Hardware Implement the communication on a Microcontroller in C (Bit-Stream Based) Demonstrate the functionality Measure the Performance, make calculations
Contact:
Description
Modeling transmission characteristics is useful for decision making in cognitive radio systems and the identification of a good model highly relays on the quality of training data. While, the quality of the training data depends on two properties, these are noiselessness and completeness. Narrowband sensing platforms are expected to always miss some radio events which makes the completeness a unique limitation of these platforms. As a solution, it is proposed to formulate the narrowband sensing as a problem of learning using incomplete training data. The primary intention is to investigate learning methods that can allow estimating statistical parameters (such as mean, variance, arrival rate etc.) of transmission characteristics of radio systems with a comparable accuracy to that of wideband sensors. For proof of concept suitable state-of-the-art modeling methods (For example linear prediction coefficient, Poisson based model, Markov Models etc.) can be used to compare the performance of suggested narrowband parameter estimation methods with that of wideband methods. Proposed method needs to be implemented using USRP2 based platform and its performance should be compared in 2.4 GHz ISM band with real-time spectrum analyzer, which will be taken as a reference wideband sensor.
Contact:
Description:
Typically, cognitive radio systems either sense the channel just before transmission or perform this task periodically in order to remain aware about the operational environment. The former method is typically known as reactive sensing, while the latter one as proactive sensing. Apart from when to sense, both methods are similar in nature, because both methods decides about the availability of sensed channel on the basis of information obtained during the sensing process. However, a channel sensed as free can become busy during the transmission of cognitive system resulting in harmful collision and unnecessary interruption for the cognitive radio system. As a solution, a predictive modeling based approach is has been proposed and has shown promising results in simulated environments. Since real-time dynamic environments offer unique challenges that can not be fairly represented using simulations, it is necessary to realize real-time demonstrators. An additional restriction is to research the use of Markov Models in this demonstrator as they are already being used successfully in other wireless application (for example Viterbi Algorithm). USRP2 based system should be used for this demonstrator. (However, it is possible to define this project only for simulated environments with slightly changed specification)
Contact:
Area: Comparative Study (software testing) Title: Evaluation of existing Programmable Logic Controller (PLC) software testing
techniques.
Description
The task involves evaluating the existing PLC testing techniques and realizing some of these techniques to support the evaluation.
Area: Comparative Study (Software testing) Title: Evaluation of existing mutant analysis techniques for software testing. Description
The task involves evaluating the existing mutant analysis techniques and realizing some of these techniques to support the evaluation.
Contact:
Description:
Having knowledge about the position of the mobile node and position of obstacles in the environment it is possible to control movements of a robot avoiding collisions. That would be especially beneficial in a scenario with many moving modes which may collide. A Software controlling and tracking movement of nodes to avoid possible collisions will be the main objective of this project work. The communication with the robot will be realized using Bluetooth technology.
Contact:
Prof. Dr. Jrgen Jasperneite juergen.jasperneite@hs-owl.de Mgr. inz Lukasz Wisniewski lukasz.wisniewski@hs-owl.de
Description:
In the hybrid wired-wireless network systems the topology may change dynamically due to the mobile nodes moving along their paths. Such changes may require some efforts to guarantee that the Quality of Service (QoS) of the critical nodes (communicating in real-time) will be satisfied. Therefore, it is vital to know in advance when such topology change happens and how the topology will look like.
Contact:
Prof. Dr. Jrgen Jasperneite juergen.jasperneite@hs-owl.de Mgr. inz Lukasz Wisniewski lukasz.wisniewski@hs-owl.de
Topic 1: Evaluating 802.1 QAT in relation with configuration requirements of industrial networks Description: 802.1 AVB based network works without engineering tool (the network
configures itself as Plug-and-Play). In industrial communication systems for the planning of IO devices and PLC programming, specialized communication planning engineering tool are used. It will be interesting to investigate how that fits together? A simulation model demonstrating possible mechanisms.
Topic 2: Investigation of redundancy options for 802.1 AVB based industrial networks Description: Evaluation of various possibilities of redundancy support for 802.1 AVB
based industrial network. What extensions may require in AVB sub specifications? A simulation model demonstrating possible mechanisms and case studies.
Topic 3: 802.1 AVB based traffic classification in relation to industrial networks Description: Ethernet AVB infrastructure to transport state of the art real time Ethernet
(RTE) protocols. Tunnelling the RTE communication through the AVB network. What are the performance impacts to the network latency? A simulation to evaluate QoS characteristics. For the functional and performance evaluation of communication networks, discrete event simulation approach is widely used (e.g OMNet ++). Which require some skills in the programming language C++ and basic knowledge in statistics.
Contact:
Prof. Dr. Jrgen Jasperneite juergen.jasperneite@hs-owl.de Mgr. inz Lukasz Wisniewski jahantzaib.imtiaz@hs-owl.de
Title:
Development of a RESTful Web Service to Access Test and Measurement Instruments with SCPI interfaces
Project Background:
For automated test and measurement tasks, modern electronic test and measurement instruments (oscilloscopes, analyzers, generators, multimeters) communicate with computer systems via a LAN, USB, or GPIB interface using Standard Commands for Programmable Instruments (SCPI). The range of application programs and instrument drivers is dominated by platform dependent, proprietary software products (e.g. LabView by National Instruments). This project intends to bring object oriented, web enabled, platform independent software technology to this kind of applications. Project Aim: The idea is, that an SCPI equipped instrument offers its capabilities as a RESTful Web service using the platform independant Java technology. Since there is a broad range of different equipment offering 3 kinds of computer interfaces (LAN, USB, or GPIB), object oriented generalization structures should be used to classify instruments, instrument commands, and hardware interfaces. Via the RESTful web interface, it must be possible to initialize an instrument, sent any SCPI command to it, and receive its response. To test the RESTful web services for various instruments, a prototype Web application should be developed, that lets a user enter SCPI commands and displays the data returned from the instrument.
References
1. 2. 3. Using XML and Java for Astronomical Instrument Control: http://www.adass.org/adass/proceedings/adass99/O8-03/ SCPI-Spezifikation: www.ivifoundation.org/docs/SCPI-99.PDF The JPIB project: Java & IEEE488 integration: http://jpib.sourceforge.net/
Contact:
Description:
The surface inspection accords a high importance in industrial production, because the product quality depends on the visual representation apart from the reliable functionality. Especially in the area of mass-production there is a general need for fast and automated surface inspection. The aim of this project is to localise errors in homogeneous and slightly granular surfaces. The algorithm has to be adapted to the surface by estimating the main spatial frequencies in the images and afterwards using this information to cluster the surface structure and topology. The defect detection is done by calculating local statistics based on the probability density without the need for training specific references.
Tool: Matlab/Simulink
Prequisites: Programming Know-how in Matlab m-language
Contact:
Scientific Assistant: M.Sc. Karl Voth karl.voth@hs-owl.de Supervisor: Prof. Dr.-Ing. Volker Lohweg volker.lohweg@hs-owl.de
Description:
The surface inspection accords a high importance in industrial production, because the product quality depends on the visual representation apart from the reliable functionality. Especially in the area of mass-production there is a general need for fast and automated surface inspection. In general products contain apart from homogeneous and slightly granular surfaces also areas with inhomogeneity (like edges, etc.). The aim of this project is to map the surface by using local support frequency based algorithms. The defect detection is executed by calculating local statistics on homogenous areas.
Tool: Matlab/Simulink
Prequisites: Programming Know-how in Matlab m-language
Contact:
Scientific Assistant: M.Sc. Karl Voth karl.voth@hs-owl.de Supervisor: Prof. Dr.-Ing. Volker Lohweg volker.lohweg@hs-owl.de
Description:
Learning a Fuzzy-Pattern-Classifer is nowadays still heuristically driven. Automatic training approaches may help improving learning situations, for example when only a small number of training examples are available. The PMFPC approach was addressed recently, but may yield results which are insufficient in the application's context. This approach is to be evaluated using various data sets to confirm its performance or exploit improvement potentials.
Contact: Scientific Assistant: M. Sc. Uwe Mnks uwe.moenks@hs-owl.de Supervisor: Prof. Dr.-Ing. Volker Lohweg volker.lohweg@hs-owl.de
Topic: Feasibility Study for the Detection of Faults on Extruded Plastic Pipes
Area: Image Processing and Pattern Recognition / Sensor Fusion / MachineDiagnosis
Start: Partner: Tool: Prequisites: how September 2010 granted Research Project Matlab/Simulink Programming know-how in Matlabs m-language, mathematical know-
Description:
The detection of faults on plastic surfaces is state of the art. Although, this task becomes challenging on curved plastic pipes. On those surfaces, different types of errors exist. Those have to be detected in first place and, secondly, be classified into different fault classes. The task of this student research project is to develop a technique for fault detection on those curved surfaces. All necessary tools like cameras, illumination, and software shall be selected. In a second step the necessary image processing algorithms shall be researched.
Contact: Scientific Assistant: Dipl.-Ing. Alexander Dicks alexander.dicks@hs-owl.de Supervisor: Prof. Dr.-Ing. Volker Lohweg volker.lohweg@hs-owl.de
should be evaluated and the appropriate testing methods for this special architecture have to be chosen. Secondly, it should then be used to test the functionality of single modules as well as the complete system functionality. If you are interested in this special topic and have a strong focus on the following interests you could be the perfect fit for us:
Software Design Programming Language C (Industrial) Ethernet: Modbus/TCP, PROFINET, EtherCAT
Contact: Alexander Dojan alexander.dojan@iosb-ina.fraunhofer.de Florian Pethig florian.pethig@iosb-ina.fraunhofer.de Supervisor: Prof. Dr. Oliver Niggemann oliver.niggemann@iosb-ina.fraunhofer.de
Contact: Sebastian Faltinski sebastian.faltinski@iosb-ina.fraunhofer.de Supervisor: Prof. Dr. Oliver Niggemann oliver.niggemann@iosb-ina.fraunhofer.de
MasterIT:ProjectworkWS2011/2012
Area:CognitiveRadio Title:StatisticalParameterEstimationUsingNarrowbandandWideband Sensors Description
Modeling transmission characteristics is useful for decision making in cognitive radio (CR) systems. The identification of a good model highly relies on the quality of training data. While, the quality of the training data depends on two properties, these are noiselessness andcompleteness.Narrowbandsensingplatformsareexpectedtoalwaysmisssomeradio eventswhichmakesthecompletenessauniquelimitationoftheseplatforms. Asasolution,itisproposedtoformulatethenarrowbandsensingasaproblemoflearning using incomplete training data. Such problems are often solved by either improving the qualityofacquireddataorbychoosingparameterestimationmethodswhicharerobustto incompleteness.Theprimaryintentionistoinvestigatelearningmethodsthatcanaccurately estimate statistical parameters (mean, variance etc. referred by i) of transmission characteristicsofprimarysignalsusingnarrowbandsensors. Tasks: 1. StateoftheartsurveyonpredictivemodelingstrategiesforCRsystems. 2. Investigatestateoftheartmethodsforlearningusingmissingdataandinvestigate theirsuitabilityfornarrowbandlearningproblems.Shortlistthreemostsuitable methods. 3. Createacoexistingscenarioandcaptureitusingrealtimespectrumanalyzerforpost processinginMATLAB/Simulink.Computestatisticalparameters(REF)ofthisdata.It willbelaterusedasreferenceforperformancecomparison. 4. Simulatenarrowbandsensingandapplyselectedmethodstocomputei.Compareit withREF.
[1] Hyoil Kim; Shin, K.G.; , "Efficient Discovery of Spectrum Opportunities with MACLayer Sensing in CognitiveRadioNetworks,"MobileComputing,IEEETransactionson,vol.7,no.5,pp.533545,May2008 [2] AlanC.Acock,;WorkingWithMissingValues,JournalofMarriageandFamily,Volume67,Issue4,pages 10121028,November2005 [3] YangYuan,;MultipleImputationforMissingValues:ConceptsandNewDevelopment,SASInstituteInc., SUGIProceedings,2000
MasterIT:ProjectworkWS2011/2012
Area:CognitiveRadios Title:PredictiveModelingforCoexistenceOptimizationusingMarkovModels Description
Typically, cognitive radio systems either sense the channel just before transmission or performthistaskperiodicallyinordertoremainawareabouttheoperationalenvironment. Theformermethodistypicallyknownasreactivesensing,whilethelatteroneasproactive sensing. Apart from when to sense, both methods are similar in nature, because both methods decide about the availability of the sensed channel on the basis of information obtainedduringthesensingprocess.However,achannelsensedasfreecanbecomebusy during the transmission of the cognitive system resulting in harmful collision and unnecessaryinterruptionofthedatatransmission. As a solution, a predictive modeling based approach has been proposed and has shown promising results in simulated environments. This project work will investigate the use of MarkovModelsforpredictivemodelingofprimaryusertrafficastheyarealreadybeingused successfullyinotherwirelessapplication. Tasks: 1. StateoftheartsurveyonpredictivemodelingstrategiesforCRsystems. 2. Theoretical analysis of Markov models to model the spectral usage. IEEE 802.11 basedWLANshouldbeselectedasprimaryuser. 3. Investigate the influence of different traffic types of selected primary users to use Markovmodels. 4. Theoreticalanalysis of computation of different probabilistic parameters of Markov models. 5. UseMATLAB/SimulinkorUSRP2basedplatformtodemonstratetheperformanceof Markovmodels.
[1] Geirhofer, S.; Lang Tong; Sadler, B.M.; , "COGNITIVE RADIOS FOR DYNAMIC SPECTRUM ACCESS Dynamic Spectrum Access in the Time Domain: Modeling and Exploiting White Space," CommunicationsMagazine,IEEE,vol.45,no.5,pp.6672,May2007 Hyoil Kim; Shin, K.G.; , "Efficient Discovery of Spectrum Opportunities with MACLayer Sensing in Cognitive Radio Networks," Mobile Computing, IEEE Transactions on , vol.7, no.5, pp.533545, May 2008
[2]