Despite the growth of advanced communication technologies, smart devices etc. the main approach to training strategic planners for crisis management (Gold Commanders) continues to be a paper-based, collective group dynamic exercise. The Pandora project has developed an advanced smart environment for the training of Gold Commanders which uses AI planning techniques to provide a crisis scenario modelled as an event network. This includes points of decision for trainees managed by automated rules from a knowledge base, behavioural modelling of the trainees, and ambient management of the environment to provide affective inputs to control and manage trainee stress. In this context, the system controls and reacts to trainee performance in relation to the events and decision points and can dynamically remodel and reconfigure the event network to respond appropriately to trainee decisions. Trainees can also be pressurised through compression of the timelines or ambient management of the multimedia information presented within the environment, causing them to make decisions under stress or with inadequate information. The environment can also represent any missing trainees within the scenario, which provides the potential to offer a completely autonomous facility for scenario design and test, and potentially a decision support facility, based on a build-up of empirical evidence from real world and training situations. In summary, the Pandora system integrates its computational intelligence, with the intelligence of the trainer and the trainees, to provide an emotionally engaging, augmented reality/virtual reality training environment for crisis managers.
Developing a smart environment for crisis management training
springer, itali, Germania
Journal of Ambient Intelligence and Humanized Computing 4 (2013): 581–590. doi:10.1007/s12652-012-0124-0
info:cnr-pdr/source/autori:Bacon, L.; MacKinnon, L.; Cesta, Amedeo; Cortellessa, Gabriella/titolo:Developing a smart environment for crisis management training/doi:10.1007/s12652-012-0124-0/rivista:Journal of Ambient Intelligence and Humanized Computing/a