Reasoning with uncertainty in the situational awareness of air targets

In combat simulations target classification and identification are very important. In this research area several studies about simulating identification have been done, most of them take a set of information like “the target is visually identified hostile” to start the simulation with. Mostly classification is not taken into account in identification problems. In this report the input consists of basic sensor data and a priori knowledge. This will be combined into information which is necessary to evaluate the situation. Based on this information the complete situational awareness is evaluated. To derive information out of sensor data, facts have to be derived in three areas, these are facts concerning position, identity and behaviour. Based on these derived facts a decision will be made about the classification and the identification of the target. Two Bayesian reasoning models were designed for the decision processes of the targets classification and identification. These models are designed as much alike as possible. An implementation was made to test the models. In the implementation temporal aspects are not taken into account but the results were promising. To conclude we conducted a literature survey to investigate the possibilities of temporal reasoning in this project.