Positions-Only Clustering on Passive Location Points of Communication Transmitters

For communication targets in a specific network, position is of great significance to recognize each of the targets. However, as one of the ways to obtain positions of communication transmitters, passive location has several problems such as uncertainty in the reception of communication signals and a low accuracy, which greatly hinder the use of recursive filtering algorithms like the extended Kalman filter. In this article, a positions-only clustering algorithm suited for passive location data of multiple communication transmitters is proposed. The most innovative idea of this method consists of introducing two types of classes, the single-element class and the multiple-element class, into the clustering process. Four clustering paths having different priorities are herein thoroughly described. For different paths, different criteria based on distance or motion continuity are available. With the help of this method, different communication targets can be recognized and their motion states and moving trails can be determined. And tests with practically collected data show that this method has the potential to relink separated groups of location points which should have belonged to the same target, as well as to distinguish between traces of two motional targets that are close to each other.