Incident Detection

For the prevention of traffic breakdowns it is necessary to identify disruptions and incidents in traffic flows as early as possible, so as to adopt the appropriate counter measures in time. As no incident is the same, sophisticated concepts are necessary to obtain a comprehensive detection of incidents and anomalies with good detection rates. Disruptions can range from faulty sensors, damaged infrastructure as well as unusual traffic flows, special events or accidents and catastrophes.

ANDATA provides a broad spectrum of ready-made software modules for the automatic detection of disruptions and anomalies in sensor signals, which can be adapted simply and quickly to the relevant operational surroundings and to the special sensors. The available detection methods range from simple threshold values to sophisticated pattern recognition procedures from the field of artificial intelligence, which can independently learn the normal course look of the sensor signal.

The definition and collation of the different descriptions of disruptions occurs in a uniform software environment. This allows expert knowledge about possible signals and disruptions to be collected, continually extended and systematically analysed.

 

Features and Advantages

  • Comprehensive software tools for the definition, administration and analysis of disruptions and anomalies in data and sensor signals.
  • Learning and continuously extendable database for the definition of potential disruptions and anomalies.
  • Universal implementation allows simple integration into arbitrary existing environments and platforms.
  • No limitation to certain kinds of sensors or information.
  • Treatment of arbitrary combinations of various sensors for checking the consistency of the information.
  • Broad spectrum of available mathematical methods from simple threshold values to sophisticated pattern recognition methods from the field of artificial intelligence.
  • Adaptive, partly self-learning methods for the identification of incidents and anomalies.
  • Numerous ready-made filtering methods and signal processing functions available for the preconditioning of the sensor signals and data.
  • Simple extension with respect to additional signal preconditioning, detection methods, statistics and learning procedures.