QAM-PPM Hybrid Modulation End-to-End Communication System Based on CNN Zhang Tonghao, Wang Xudong*, Wu Nan Information Science and Technology College, Dalian Maritime University, Dalian 116026, China Abstract In order to optimize the structure and performance of the quadrature amplitude modulation (QAM) and pulse position modulation (PPM) hybrid modulation system applied to visible light communication, a hybrid modulation end-to-end communication system based on convolutional neural network (CNN) is proposed. This scheme uses the designed loss function to train the network in multiple stages to realize QAM modulation and PPM modulation respectively, then the two are combined to realize hybrid modulation. In terms of demodulation, in order to improve the accuracy of pulse recognition and reduce the complexity of calculation, a method for pulse recognition of the received signal by changing the kernel size of CNN is proposed. The simulation results show that under the additive white Gaussian noise channel and the Rayleigh fading channel, the proposed technical scheme shows fine generalization ability for the hybrid modulation method with different pulse time slots and modulation levels, when the symbol error rate is 10-3, the e
