The application of feed-forward connectionist models to ESM bearing estimation using signal amplitude

Bearing estimation is a central task for electronic support measures (ESM). The bearing or direction of arrival (DOA) of a signal is a powerful sorting parameter even without yielding identification information. Connectionist models (CM) are networks of simple processing elements that perform multidimensional classification or function mappings. The bearing estimation problem requires a multidimensional function that takes information from an antenna array and computes the DOA. The authors show that CMs can be trained to perform this task. The authors use an azimuth array of six identical elements spaced at 60° to measure the signal amplitude. The data for each antenna was measured in an anechoic chamber at frequencies from 2 to 18 GHz at 2 GHz intervals