A Situation Analysis Decision Support System Based on Dynamic Object Oriented Bayesian Networks

This paper proposes a situation analysis decision support system (SADSS) for safety of safety-critical systems where the operators are stressed by the task of understanding what is going on in the situation. The proposed SADSS is developed based on a new model-driven engineering approach for hazardous situations modeling based on dynamic object oriented Bayesian networks to reduce the complexity of the decision-making process by aiding operators’ cognitive activities. The SADSS includes four major elements: a situation data collection based on observable variables such as sensors, a situation knowledgebase which consists of dynamic object oriented Bayesian networks to model hazardous situations, a situation analysis which shows the current state of hazardous situations based on risk concept and possible near future state, and a humancomputer interface. Finally two evaluation methods for partial and full validation of SADSS are presented