Universität Duisburg-Essen Home

How to find us
News / Conferences
External Links


Studies / Lectures
Final theses /
Project works: Topics

Heinz-Luck-Fire Lab

AUBE '21 / SUPDET 2021


Electrical and Electronics

Alumni Engineering Sciences
Engineering Sciences
Universität Duisburg-Essen

Mobile radio:
Smart Antennas
Adaptive MIMO
     communication systems

Filter bank multicarrier
    transmission systems

Adaptive multicarrier

Wideband GHz and THz
     radar sensors

Fire detection:
Automatic Fire Detection
Test devices for
     dust and water fog tests

     Aerosol Charactersiation


Fire Detection

Stand: 08.08.2016
Universität Duisburg-Essen, Campus Duisburg
  • German damage insurers report direct insured losses due to fire of several billion euros yearly. Compared to the losses in industry and commerce, about 20% of the losses are recorded in residential environments. Adding consequential losses due to e.g. business interruption and losses covered by other types of insurances like liability insurance, the total amount of damages caused by fires is probably two or even three times higher.

    These massive economical losses stress the importance of automatic fire detection and suppression systems for residential, commercial and industrial environments. A key aspect of any automatic fire detection and suppression systems is a reliable signal detection, which can be defined as a special case of a “signal detection in noise”. The technical conditions to be met by cutting-edge signal processing approaches can be described for instance by the following requirements:

    • The signal detection must be done as fast as possible, as the costs involved strongly depend on the detection time.
    • The false alarm rate shall be as low as possible, as the actions triggered by an alarm can cause extreme high costs (e.g. a forced landing of an airplane or the unwanted release of a fire suppression system). These costs are completely unacceptable in the case of false-alarms.
    • The signal to be detected is generally not known in all details, as it is triggered by an unwanted or hazardous situation and affected by numerous parameters.

    The problems related to the given requirements have not been completely solved so far. There are many fully functional fire detection systems with very good fire detection characteristics, but unfortunately the complex task of avoiding false alarms is not completely addressed. In addition, with increasing number of fire detection systems the overall number of false alarms also increases. Following approaches are e.g. pursued to reduce the false alarm rate:

    • The implementation of different sensors for fire detection, as the combination of several physical phenomena consolidates the decision quality.
    • A more sophisticated signal processing of single- or multi-criteria detection systems, as several conclusions about the signal source may be drawn from particular signal characteristics.
    • Improved sensor technology using polarimetric characteristics.
    • Novel methods for the test of fire detection systems in scenarios with non-fire aerosols.

    The technological progress in the field of sensors and microcomputers pave the way for more sophisticated signal processing approaches on energy-saving and low-cost systems. Usually there is no mathematically manageable method for the determination of the detectivity of a detector (i.e. its efficiency considering all risks associated with false alarms and missed alarms probabilities). At least the detection characteristics of a fire detector in the case of a fire can be evaluated in a relatively short time on the basis of well standardised test-fires performed in the Heinz-Luck-Fire-Lab of the Chair of Communication Systems.

    However, far more difficult is the determination of the behaviour of a detector in non-fire situations, e.g. by means of its false-alarm rate. These analyses get more and more challenging with increasing robustness of detectors against false-alarms. Target of modern detection systems is to show not more than one false alarm per detector unit and year, without negatively affecting the fire detection characteristics.

    Research topics:

    • Detection algorithms / Simulation methods
    • Polarised aerosol characterisation
    • Sensor modelling
    • Test methods for fire detection equipment
    • Video-based fire detection