ROTHR Mode-Linking Using Neural Networks
Final technical rept. Jul 1995-Dec 1996
RIVERSIDE RESEARCH INST LEXINGTON MA BOSTON RESEARCH OFFICE
Pagination or Media Count:
A novel concept using neural network technology was employed to find the correct linkage of ionospheic modes in a multi-mode multi-target environment given a model of the ionosphere. The technology employed consists of a self-organizing feature map SOFM to provide separation of multiple targets containing multiple modes followed by an Adaptive Resonance Theory Map ARTMAP to perform spatial pattern recognition of ionospheric mode patterns and to link them to a single ground range. The ARTMAP recognition system is trained on the Coordinate Registration CR tables produced by ROTHR. The current CR tables are updated every 12 minutes, thus providing a real-time assessment of the ionosphere. The exactness of the mode-linking procedure consequently depends on the accuracy of the Coordinate Registration tables.
- Atmospheric Physics
- Active and Passive Radar Detection and Equipment