Signal and Information Processing for Resource-limited Applications

A large number of our measurement applications use indirect measuring principles for the sensory acquisition of the quantity to be measured. After acquiring the measured variable by the sensor and the measurement system, the information about the quantity must be determined from the measurement signals.

We are working on the development of signal processing methods for the evaluation of measurement signals in energy critical applications. The research focus lies in the development of model-based evaluation approaches, which can be optimally adapted to the measurement task and executed on a computer system with frequently limited processing power.

Our main application areas are in the field of power engineering, heavy industry and the automotive industry. Examples can be found in the tomographic condition monitoring of transport processes in heavy industry, the condition monitoring of transmission lines, or the sensory monitoring of internal combustion engines.

Many of our applications are characterized by a complex character that influences the design of signal processing and is reflected in the algorithms. These challenges include for example:

• Nonlinear sensor and measurement models, which often need to be described by PDE models.
• Under-determined  and ambiguous relations between the measurements and the quantities of interest.
• Cross influences on the measurement process, which must be taken into account in the evaluation.

The holistic physical and statistical modeling of the sensor effect and the measurement system is therefore a central component of our research work. Based on these models, measurement information with prior knowledge and additional information can be optimally used for the design of estimation methods.

The tools of our work are in the field of simulation and modeling, technical signal analysis and statistical signal processing. The breadth of these topics and the variety of applications are also reflected in the expertise of our employees.

In connection with our applications, the transfer of optimized signal processing methods to resource-limited process systems is another important issue. In this context, we also deal with aspects of the uncertainty quantification, as well as the combined design of measurement system and signal evaluation system.

Selected research activities and results:

• Tomographic monitoring of funding processes
• Condition monitoring on overhead lines
• Predictive maintenance for internal combustion engines