Prediction of frequency parameters in short wave radio communications based on chaos and neural networks

In order to improve the reliability of short-wave communication, a hybrid method of prediction based on chaos phase reconstruction and neural networks is proposed and used in frequency parameters prediction. We use the chaos method to reconstruct attractors in phase spaces, fit the attractors' global map by multi-layer feedforward neural networks, and thus construct a hybrid model of prediction. Experimental results show that the hybrid model can achieve good results in predicting frequency parameters of short-wave communications such as foF2, and has promising applications. We also show the efficiency of a noise-suppressing method based on single value decomposition (SVD).