Research On Self-interference Cancellation Of Co-time And Cofrequency Full-duplex OFDM System

With the development of communication technology, demand of the improvement of the spectral efficiency and the transmission speed is becoming more and more serious in wireless communication system. Compared to half-duplex mode, full-duplex communication technology can avoid the spectrum resource waste, thus possible to realize the channel capacity of the communication system doubled. However, the strong interference signal from the local transmitter is the biggest obstacle to the full-duplex communication system. Research on the self-interference cancellation is one of the most important parts of the system. Based on these analyzes, this article will focus on selfinterference cancellation in analog and digital domain, and verify the effectiveness of the system.Firstly, study the basis of full-duplex communication system and system model. The paper introduced co-time and co-frequency full-duplex communication system and self-interference channel model. On the condition of co-time and co-frequency fullduplex communication, the communication channel model is the Rayleigh fading model, and the self-interference channel is the Rice model with large K factor.Secondly, study the self-interference cancellation in analog domain. The paper reduced the self-interference by cancelling the reconstructed self-interference. The system reduces the complexity of reconstructing self-interference by using the product in frequency domain instead of convolution in OFDM system by IFFT. Experiments on the condition of high INR shows that the performance of LS and MMSE on selfinterference channel estimation are similar. The system can reduce as much as 32 d B self-interference.Thirdly, study the high dynamic range analog to digital converter algorithm. Analyze the necessity of improving the dynamic range analog to digital converter, while the SINR is low. Proposed a cascade form ADC algorithm based on the compensation of the quantization error of ADC, and verify the effectiveness of the algorithm. The effective dynamic range of chosen AD and DA should not be less than 50 d B and 75 d B. Simulation results show that the algorithm can improve 20 d B of dynamic range.Finally, study the self-interference cancellation in digital domain. In the process of reconstructing the digital self-interference signal, improve the original channel estimation algorithm. Considering the self-interference channel state information changes gently, self-interference channel estimation in which frequency domain least square algorithm is used is done in the training mode of half-duplex. Simulation results show that it can reduce 15 d B self-interference more than the conventional channel estimation algorithm. Study the amount of self-interference cancellation of the whole system, as the system can reduce 68 d B self-interference, with 36 d B of digital selfinterference cancellation. The self-interference can be reduced effectively while the SINR is more than-55 d B.In a word, the paper studies the active self-interference cancellation on the condition of co-time and co-frequency full-duplex communication system. Proposed a high dynamic range analog to digital converter algorithm and verify the effectiveness of the algorithm. The effectiveness of self-interference cancellation and the whole system are verified by the experiments on the index of the amount of the self-interference cancellation and BER.