Target identification with dynamic hybrid Bayesian networks

Sampsa K. Hautaniemi, Petri T. Korpisaari, Jukka P. P. Saarinen, Sebastiano B. Serpico
2001 Image and Signal Processing for Remote Sensing VI  
The continuous growth of data has created a demand for better data fusion algorithms. In this study we have used a method called Bayesian networks to answer the demand. The reason why Bayesian networks are used in wide range of applications is that modelling with Bayesian networks offers easy and straightforward representation for combining a priori knowledge with the observations. Another reason for growing use of the Bayesian networks is that Bayesian networks can combine attributes having
more » ... ferent dimensions. In addition to the quite well-known theory of discrete and continuous Bayesian networks, we introduce a reasoning scheme to the hybrid Bayesian networks. The reasoning method used is based on polytree algorithm. Our aim is to show how to apply the hybrid Bayesian networks to identification. Also one method to achieve dynamic features is discussed. We have simulated dynamic hybrid Bayesian networks in order to identify aircraft in noisy environment. Keywords: Bayesian networks, attribute association, type identification. Identification, or hostility, of the target is almost impossible to determine using only its kinematic components (such as position, velocity, acceleration). However, it is known that every aircraft has its own specialities such as the form of the wings, the number of the engines, an ability to use only certain band of frequencies etc. In this study these beforehand known (a priori) quantities are referred as attributes. The problem of using attributes is that dimensions of the attribute observations can be almost anything, e.g. frequency is given in Hertz, IFFN has no dimension, exhaust fumes are measured in kelvins etc. Therefore, the association of attributes is difficult. In addition, the association of somehow combined attributes and kinematic observations is even more difficult. There are proposed only few theories, which can be used in attribute association. Bayesian networks are one of them 1,5 . The purpose of this paper is to represent theory of the hybrid Bayesian networks and apply it to an identification system. The order of this paper is as follows. First, we define a few preliminary definitions. Then the theories of discrete and continuous Bayesian networks are briefly introduced. The theory of the hybrid Bayesian network is represented in detail. After theoretical discussion we show the modification needed to Bayesian networks in order to solve identification problem. Simulations of dynamic hybrid Bayesian networks in noisy environment are illustrated in Chapter 7.
doi:10.1117/12.413885 fatcat:4jbhakig3jat5cxqrjojhod35m