Neural network beamformer for narrow-band HF transmission

The multi-layer perceptron (MLP), an artificial neural network, is applied to adaptive beamforming over narrow-band HF channels. An HF data transmission system is simulated, incorporating QPSK modulation, demodulation, space diversity channel model and adaptive beamforming. The simulated diversity channels are characterised by Rayleigh fading and have time-varying characteristics. MLP beamformer has the ability to learn the statistical behaviour of the channels and to correct the distortion they introduce to the transmitted signals. Simulation results for the MLP beamformer using the back-propagation algorithm and for the conventional beamformer using the least-mean square (LMS) algorithm are reported and evaluated. Improved performance is exhibited by MLP beamforming techniques