Guido Dartmann

Latest Publications

Dziubany, M., Machhamer, R., Laux, H., Schmeink, A., Gollmer, K.-U., Burger, G. and Dartmann, G.: Machine Learning Based Indoor Localization Using a Representative k-Nearest-Neighbor Classifier on a Low-Cost IoT-Hardware, in Proceedings of European Signal Processing Conference (EUSIPCO) , p.p. 2050-2054 Sep/2018 , 10.23919/EUSIPCO.2018.8553155

Pallasch, C., Peitz, A., Herfs, W., Schmeink, A. and Dartmann, G.: Novel approach for wireless commissioning and assisted process development based on Bluetooth Low Energy, in 23rd International Conference on Emerging Technologies and Factory Automation , IEEE, Jul/2018 , 10.1109/ETFA.2018.8502455

Vieting, P., de Lamare, R. C. , Dartmann, G. and Schmeink, A.: An Adaptive Learning Approach to Parameter Estimation for Hybrid Petri Nets in Systems Biology, in IEEE Statistical Signal Processing Workshop (SSP) , p.p. 543-547 Jun/2018 , 10.1109/SSP.2018.8450824

Laux, H., Bytyn, A.Ascheid, G., Schmeink, A., Karabulut Kurt, G. and Dartmann, G.: Learning-Based Indoor Localization for Industrial Applications, in Workshop on Sensor Data Fusion and Machine Learning for next Generation of Cyber-Physical-Systems in conjunction with ACM International Conference on Computing Frontiers 2018 , p.p. 355-362 , ACM New York, NY, May/2018 , 10.1145/3203217.3203227

Zechendorf, E., Vaßen, P., Zhang, J., Hallawa, A., Schuerholz, T., Martincuks, A. , Krenkel, O. , Müller-Newen, G. , Simon, T.-P., Marx, G., Ascheid, G., Schmeink, A., Dartmann, G., Thiemermann, C. and Martin, L.: Heparan Sulfate Induces Necroptosis in Murine Cardiomyocytes – A Medical-In-Silico Approach Using Machine Learning, In Frontiers in Immunology Feb/2018 , accepted for publication , 10.3389/fimmu.2018.00393

Full list

Research Interests

My research is at the interface between electrical engineering and computer science. I work on data analytics, signal processing and security in communication systems. Many concepts and ideas for secure cyber-physical systems require interdisciplinary knowledge in various areas such as digital signal processing, data analytics, machine learning, optimization, systems theory, communication technology, distributed systems and software security, which I have applied in the last years in various research questions. 

The Internet of Things as the basis for Industry 4.0 is omnipresent and provides science and society with new challenges. The term cyberphysical systems is often used in the media to make it clear that technology will merge with our environment in the future. In addition to risks such as threats to information security, which are a central topic of my research, this networking and digitization can also open up great opportunities to meet global challenges. Examples of this are: improved environmental protection through continuous collection of environmental data with sensor systems or more efficient and flexible production in a factory of networked and mobile robots. Classical concepts of digital signal processing must be further developed for safe cyber-physical systems.

Due to their distributed and decentralized nature, new distributed and efficient optimization methods and algorithms must be developed for cyber-physical systems. This decentralized structure also opens up various opportunities to attack IT security. 

Besides the decentralized structure of those systems, the analysis of (sensor) data is another challenge. The mere recording and administration of data has no additional profit for a company. The analysis of the data and the resulting inference based on such process data has a large potential for future technical systems. This data analysis in turn offers opportunities to detect IT attacks at an early stage and reduces the negative consequences of these attacks. In addition to new communication technologies and data management platforms, new efficient algorithms must be developed to analyze this data.

Research Area 1 - Platforms and Software for the Internet of Things (Enabling Technology): 

IoT platform: Together with IoT expert group of the national Digital Summit, we have developed a generic IoT gateway. This gateway is characterized by low costs and flexibility. An important aspect of this gateway is the development of a generic hardware interface to various components in industrial applications. Another aspect is the connection of additional sensors. Radio-based sensors are interesting here because of their flexibility. The gateway should also be a basis for industrial projects.

Aspects of data analysis: Together with the expert group Internet of Things I am working on a monitoring system for M2M cyber security. The goal here is to develop a certified IoT hardware platform that can serve as an SIEM-Sensor (SIEM: Security Information and Event Management). This platform serves as a honeypot in IoT systems and is intended to record data traffic and forward it for analysis and anomaly detection (see F3).

Adaptive software interfaces:  Besides the hardware platform, a semantic software interface is another important aspect of this project. Due to the diversity of different connected components, aspects of the respective hardware must be abstracted. Self-learning software can learn new components and in future automatically generate test code for the adaptation of the interfaces.

Research Area 2 - Real-time Data Analysis: 

Real-time data analysis for multi-agent systems: Many cyber-physical systems are so-called multi-agent systems, i.e. they consist of several independent agents with a common goal. These agents are usually not controlled by a central unit. Typical examples are micro-satellites or swarms of drones. For example, to hold a formation, the agents must transmit their local information (e.g. state vector, distance) to their neighbors. The agents have to communicate via radio and thus deal very efficiently with the available resources, i.e. communicate only when it is necessary. Therefore, learning methods that can learn and predict the system behavior of the entire agent system are of interest here. Furthermore, algorithms for localization and learning and estimating future system states must be distributed and extremely efficient, since most multi-agent systems have only a limited energy budget. 

Learning Methods for Indoor Localization: For many applications in the field of storage and production technology, inexpensive and reliable localization in the indoor domain is very relevant. We are currently investigating methods that can determine fingerprints of sound signals and thus continuously learn a position in space. With these methods, inexpensive microphones could be used for localization. The microphones could, for example, be installed in the IoT Octopus developed by the IoT expert group. 

Deep learning in the automotive sector: One challenge in the automotive sector is the development of machine learning processes for autonomous driving. The range of these methods is wide and there exits now powerful hardware platforms that can implement such machine learning methods in real time. To achieve real-time capability, the optimization of these algorithms is elementary. In the PARIS project parallelization methods are developed to optimize these algorithms. The next goal in autonomous driving is fleet intelligence. To achieve this goal, local learning methods must be combined with global optimization methods.

Research area 3 - Big- and Small Data Analytics:

Big Data Analytics: In addition to hardware and software, the integration of an analytics tool chain for processing data is an important aspect for applications with many sensors and data. This includes e.g. classification, regression and logistic regression of measurement data. For this purpose, machine learning concepts are to be provided in software blocks and integrated into the hardware software platform of research area 1. Additional data can be recorded by the gateways, e.g. in a production facility. With this data, processes can be optimized and new business models can be generated. The process data can be collected and further processed in an edge cloud system. For this purpose, algorithms must be developed in the research field 2. In addition to the optimization of production processes, data analysis is important for an anomaly detection in order to detect IT attacks at an early stage. Anomaly detection requires suitable models of the overall system and continuous recording of all system states. An important aspect of this modeling is to find the necessary depth of detail. The following applies here: As simple as possible but also as precise as necessary. 

Big Data Analytics (medical technology): In this project we want to develop adaptive models for the diagnosis of various diseases in intensive care units. These models help young physicians to better understand the complex processes involved in sepsis, for example. The system should be able to predict different diseases using adaptive models and measurement data. The system continuously records medical measurement data from the telemedicine system in a large database. On the basis of this constantly growing database, models for specific diseases can be trained. These models should be valid for a large number of patients as well as individually optimized for certain classes of patients. The goal is to develop individualized patient models for intensive care disease processes.  Both neural networks and stochastic Petri-nets are suitable models. 

Small Data Analytics (data analysis in systems biology): A major challenge in medicine is machine learning with small amounts of data (low data problem), because the generation of data for drug research is very expensive. For this reason, we want to develop so-called bio-in-silico processes for model-based computer-assisted drug optimization in cooperation with physicians. Such methods reduce the number of necessary laboratory tests and could make drug development more cost-effective. Continuous Petri networks, which are already frequently used for the modelling of systems biology processes, are suitable models.  Here we investigate algorithms that learn the parameters of given Petri nets (structure known) as well as concepts that learn the entire structure of these Petri nets in addition to the parameters. In contrast to neural networks, models based on Petri-nets are not black-box models. Such white-box models are more accepted in medicine and also allow to integrate expert knowledge.

Research Area 4 - Privacy and Security for Distributed Systems and IoT: 

Without security, cyber-physical systems are never accepted in industry. A future cyber-physical system must be able to independently detect possible attacks with the help of data analytics, for example. 

Privacy: On the other hand, cyber-physical systems can also endanger the privacy of private individuals. The Internet of Things collects and merges data from thousands of sensors. Increasingly powerful machine learning processes have made it increasingly easier to merge data and deduce information. New concepts are needed here to ensure tap-proof communication and concepts for anonymization for sensitive personal data.

Physical-Layer Secruity: Radio-based systems have become more and more interesting for industrial applications due to their flexibility. In traffic (autonomous driving) communication is only possible via radio. Many of those systems are embedded and do not have complex encryption mechanisms. Due to the distributed nature of the radio channel, it is an easy target for attackers. That's why I've been researching concepts to achieve IT security in layers 1 and 2 for some years now. Here I cooperate with my colleague Prof Gunes Karabulut Kurt of ITU Istanbul. 

 More details of my research group at University of Applied Sciences Trier.