Target Threat Assessment based on Ensembles of Multi-Criteria Decision Making Methods (Poster)

Aerial target threat assessment is an important but stressful task carried out by operators. The work involves determining the threat of aerial targets based on various types of information about the target, such as kinematics, specifications, intelligence reports, and others. The success of threat assessment lies in the reliable fusion of this information. The recent progress in technology allows for the development of more complex decision making methodologies to take into account all types of information, latency, environments, changes in priority, and inconsistencies in the quality and availability of data sources. It also allows for the integration between different decision making methodologies in order to take advantage of the strengths of each independent methodology for different applications. Multi-criteria decision making (MCDM) methods provide a structured and intuitive framework for incorporating all available target information to evaluate the target threat. However, threat assessment is a complex process where there is not one MCDM method that works in all situations. The objective of this work is to develop a robust system that uses various MCDM methods as an ensemble (i.e. an EMCDM technique) and machine learning techniques to assess the threat of aerial targets. Experiments involving threat assessment on simulated targets indicate the benefit of using EMCDM techniques instead of using individual MCDM methods for performing high-level information fusion to assign threat values.