Neural Network Based Demodulator for Known Binary Baseband Signal
Weidong Cheng, Wang Tianbao, Zhan Wen · 2009
In this paper, a novel neural network signal demodulator model for known binary baseband signal is proposed. Simulation results shown that the performance of the neural network demodulator proposed outperforms the performances of the correlator model and the matched filter model. To achieve the same bit error probability, the input signal to noise ratio of the proposed model is 3dB lower than the conventional models. It is important to explore and to develop the new known signal demodulator based on neural network to enhance the quality of the communication system. And the method of changing the time sequential signal operation to parallel operation is highly valuable for developing new signal detection method and theory.