Experiments of direction finder by RBF neural network with post processing

Evaluation tests of a direction finder, whose antenna aperture consists of 4 monopoles on the circumference of a 0.4/spl lambda/ diameter circle, has been carried out. An RBF neural network with a new post processing system is used for the DOA estimation algorithm. The estimation error by the RBF neural network has reduced to 1/10 compared with MUSIC.