UFMC-MIMO系统中串扰和HPA非线性失真补偿方案

The Compensation of Crosstalk and HPA Nonlinearity in UFMC-MIMO Systems

为实现通用多载波形(UFMC)和多输入多输出(MIMO)相结合的可靠的通信系统,需要一种能高效补偿由MIMO系统中非线性高功率放大器(HPA)引起的失真和非线性串扰带来的综合问题的技术。本文采用一种基于交叉神经网络(CONN)的非线性补偿器来修正非线性串扰和HPA的非线性失真,并且改进了神经网络的输入层,采用抽头延迟线技术进一步解决了记忆性非线性串扰的影响。补偿器利用反向传播的神经网络的拟合特性,采用Levenberg Marquardt(LM)训练算法训练拟合出受非线性串扰影响的HPA逆函数,将其置于HPA后以达到自适应补偿的作用。通过仿真试验证明了该补偿方案能很好地补偿串扰和非线性失真对系统性能的影响,且对于加入记忆性串扰的非线性失真具有良好的补偿特性。

In order to achieve a reliable communication system that combines Universal Filtered Multi-Carrier(UFMC)and Multiple-Input Multiple-Output(MIMO),a good method that can efficiently compensate the nonlinear crosstalk caused by the MIMO system and the nonlinearity caused by the high power amplifier(HPA)is required.A nonlinear compensator is used to correct the nonlinear crosstalk and HPA nonlinearity based on crossover neural network(CONN),and the input layer of the neural network is improved to further solve the memory effects of crosstalk by adopting the tapped delay line.The compensator applies the back-propagation neural network in which response can approximate the inverse transfer functions of HPA nonlinearity considering the nonlinear crosstalk by using the Levenberg Marquardt(LM)training algorithm,which places it after the HPA to achieve the effect of adaptive compensation.Simulation results show that the compensation scheme is able to significantly make corrections of the combined HPA nonlinearity and the nonlinear crosstalk on the system performance.At the same time,it has good compensation characteristics for nonlinearity which suffers the memory crosstalk.