Traffic detection

The following applications arise in the field of traffic detection:

  • Machine Learning and pattern recognition procedures are the methods of choice within traffic and environmental prediction, allowing an efficient and effective identification and forecast of the relevant traffic situations based on arbitrary sensors and information.
  • Virtual sensors can convert existing sensor data into the according traffic values for uniform controls, allowing the best possible usage and integration of already existing and installed sensors. Virtual sensors can also be used even when only measurements with time restrictions are available.
  • Different sensor technologies (e.g. floating car data in combination with stationary sensors) can be combined in the form of sensor fusion with the help of virtual sensors. This allows the traffic situation to be assessed much more precisely, exploiting the advantages of the relevant sensor technologies or compensating their disadvantages.
  • The requirements management and the system rating make it possible to derive precise specifications of the necessary sensors to detect the relevant traffic situations for the control. With the help of special mathematical procedures an automated identification of the best possible sensors for a certain control problem can be achieved. This ensures that only the necessary values are measured based on the control requirements and assuring economically reasonable traffic sensor equipment.
  • Incident detection can be used to check sensors as regards their consistency and plausibility. Hereby incidents and potential sensor failures can be treated within a common software environment.
  • Incident detection can be used to identify new traffic situations, which have not yet been considered in the calibration of the control but for which the control should be adapted to improve control performance.