关键词: 光正交频分复用, 分层索引调制, 机器学习

Abstract: To solve the problem that the transmission rate of traditional DC-biased optical orthogonal frequency division multiplexing (OFDM) with index modulation is not ideal, a DC-biased optical OFDM with layered index modulation scheme is proposed by introducing a layering technique and using the rotating constellation mode for layer discrimination. After elaborating the layer mapping principle of the scheme, the theoretical bit error rate under the maximum likelihood criterion is derived, and the performance is compared with the traditional index modulation system. The results show that the proposed scheme not only improves the transmission rate, but also improves the bit error performance. Compared with the traditional index modulation system (8,6,2), the transmission rate of the (8,2,3,2) layered system is increased by 20%, and the signal-to-noise ratio of the latter is improved by about 3dB when the weak turbulence and bit error rate are 10-4. Secondly, to solve the problem of high complexity of the maximum likelihood detection algorithm, a low-complexity detector was proposed according to the characteristics of layered signals combined with backpropagation neural network. Compared with the maximum likelihood detection algorithm, the detector only loses about 1.2dB of bit error performance in the (8,1,2,2) layered system, but its complexity is reduced by 75%. Finally, the feasibility of the proposed scheme is further verified by using the indoor wireless optical communication experimental system.

Key words: orthogonal frequency division multiplexing, layered with index modulation, machine learning