Reduction of false alarmsthrough machine learning of environmental noise:

  1. Clutter (or naturally existing noise in the environment) affects the readouts of the chemical sensors, which may result in a higher number of false alarms. The aspect of clutter and its influence on sensor readouts has not been given much attention in state-of-the-art systems.
  2. EU-SENSE applies machine-learning based anomaly detection method for reduction of false alarms.

Improvement of situational awarenessthrough dispersion modelling of the hazardous substances:

  1. The EU-SENSE system will apply inverse modelling that will allow for quick and precise threat source location. The algorithm will estimate potential threat source locations as soon as the system raises an alarm and the network collects first positive readouts. Subsequently, the estimation will be improved based on the incoming sensors measurements.
  2. The system will also be able to predict possible dispersion routes of the hazardous substance over time by the application of ensemble forward modelling. This information will facilitate the application of appropriate countermeasures and speed up the identification of safe evacuation routes.

Live demonstrationof fully operational solution in professional Polish Firefighter training facility in Zamczysko Nowe:

  1. The EU-SENSE system, including all functional software components and heterogeneous sensor nodes will be tested with the use of chemical simulants.
  2. The demonstration will be performed in realistic operational conditions with the involvement of firefighters including specialised chemical rescue unit.
  3. Prior to the demonstration, a comprehensive training session will be held, where the involved personnel will have the opportunity to familiarize with the equipment and the system.