Clustering Methods for Multi-sensor Data Fusion
Han Liu, Zhang Lei · 2012
In the network-centric warfare, the complete battlefield situation information and attack information for weapons can be collected and calculated by multiple sensor platforms, and the results can be distributed to the command and control systems and weapon platforms to complete target attack tasks. Track initiation is the primary problem in multi-source multi-target track processing, correct track initiation can not only reduce the calculation burden of post-processing in multi-target tracking, but also be meaningful to the discovery of interfered interesting target. Centralized algorithm first fuses the raw sensor data, and then starts the track tracing with fused points, weight the clustered sample points to fuse sample points. Clustering could be generalized from the active radar side and the passive radar side. Further more, effects caused by obstacle constraints are also discussed.