The CHARADE project (Esprit 6095) developed a system that enables firemen to save time spent on evaluating forest fire emergencies and resource engagement, thus saving lives, equipment and money.
However the system has not been tested during a real fire situation. Potential users therefore are not yet confident of its performance capability in large scale situations or its suitability to their operational needs (e.g. 24 hours a day, 7 days a week).
The French South East Inter-Regional Centre of Operational Coordination of the Civil Security (CIRCOSC) and the French International Training Centre of the Civil Security (CIFSC) have agreed to experiment with a tailored CHARADE system in shadow mode.
The aims of the CARICA project will be to demonstrate in an experiment the suitability and the benefits of the case-based reasoning (CBR) technology introduced in CHARADE, and to show its ability to support intervention planning. This project will adapt the CHARADE demonstrator for the French organisation and the results evaluated, to enable enhancement and tuning from a real context. The objective will be to ensure that it will be ready to run in a real mode for the 1996 fire campaign.
The CARICA system will provide firemen with a tool to populate a "Forest fires interventions" database which will be accessed by trainees to increase their knowledge by auto education, by firemen to study or debrief forest fires intervention, and by trainers to build exercises based on real emergencies.
If the system demonstrates its capacity in a real mode in 1996, it will be deployed in the South of France "Départements". It will also be used at the CIFSC where firemen, coming from various Mediterranean countries, will learn how to use it and be able to export the knowledge back to their respective countries.
To achieve operational and training systems customisation, the CARICA project will:
- adapt and tune the CHARADE prototype for use in a French context,
- provide a tool to enable firemen to easily feed their "Forest fire interventions" database,
- verify and tune CBR machine learning algorithms,
- improve fire assessment functions grounded in the real world,
- prepare a support training system based on auto education, interventions studies and debriefing.
Funding SchemeCSC - Cost-sharing contracts