The minimum description length criterion applied to emitter number detection and pulse classification

In this paper, we provide a method to classify incoming radar pulses according to their emitters when the number and nature of the emitters are unknown. The method is based on the minimum description length (MDL) criterion. We derive new description length measures according to different noise characteristics. The novelity of our approach is that the MDL criterion is used to obtain both the number of emitters and the optimal clustering of the received pulses. The performance of this new clustering principle is encouraging, by comparison with some known non-parametric methods.

[1]  Donald W. Bouldin,et al.  A Cluster Separation Measure , 1979, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  Richard C. Dubes,et al.  Cluster Analysis and Related Issues , 1993, Handbook of Pattern Recognition and Computer Vision.

[3]  Jun Zhang,et al.  A Model-Fitting Approach to Cluster Validation with Application to Stochastic Model-Based Image Segmentation , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  J. Rissanen,et al.  Modeling By Shortest Data Description* , 1978, Autom..