Adaptive Level of Automation for risk management

摘要

Abstract: Robots will know a huge deployment in the near future. Several research works are related with human-robot interaction and with cooperation between robots. Other studies aim to bring them capabilities of self-organization which allow them quick and appropriated autonomous answers to quite simple events. That endows the swarm of robots with interesting capabilities of resilience and adaptation. But these objectives will be limited especially when the individual and collective robot abilities will be insufficient to cope with complex demands of the environment. The only solution is therefore to ask the human help. Nevertheless the behavior of a self-organized swarm of robots is difficult to predict by the human, due to his/her difficulty to build and especially to update on line a precise representation of the situation. A first way to cope with this problem is to analyze the robots’ tasks according to several levels of activity: a strategic level in which the global objectives are decided and the whole planned mission, the tactical level for applying and updating the mission regarding new events or new objectives and the operational level for implementing decisions. A complementary way is to increase cooperative capabilities of robots and to adapt robots’ tasks in order to find the best organization between robots and humans. Our proposal is to answer this problematic by studying Adaptive Level of Automation in the domain of crisis management and firefighting.