Dynamic Bayesian networks for situational awareness in the presence of noisy data

In this paper we present a decision support system for situational awareness of combat vessels. The system provides a human operator with advice on classification and identification of air targets approaching the ship. The input of the system consists of possibly noisy sensor information. To combine data of several sources and to come to the best possible classification in the light of uncertain and incomplete information we implemented our system using dynamic Bayesian networks. Test show that the system can correctly classify air targets. Taking temporal information into account makes the system quite robust in the face of noise in the sensor data.