Spread spectrum digital signal synchronization using neural networks

The authors investigate whether an artificial neural network (ANN) can be trained to estimate the phase of the signal carrier frequency and acquire the timing of the spreading code sequence, only from samples of the received signal. They present an overview of the current work, the results obtained, and the future developments expected in this area. Very secure digital radio communication can be obtained. The frequency diversity inherent in the large bandwidth virtually eliminates the effect of drop-out due to multipath signal cancellation. A two or three-layer perceptron artificial neural network was used for synchronizing both the phase of the carrier frequency and the spreading code sequence. The network was trained with random samples of the code where the known phase was used to correct the network parameters. It was shown that the network could be trained to recognize the phase of a test sample after training with about 10000 presentations of the code. Results are given showing the performance of the system when presented with random samples of noisy signals