Radio Frequency Fonlinear Analysis And Compensation Algorithm Research

With the development of science and technology,the deepening of software defined radio theories,the modern electromagnetic environment has become more complex and more criteria have been requested,which has made it an essential element in wireless communication system study on broadband digital wireless receiver with high dynamic range and high sensitivity.Broadband digital wireless receiver has strict requirements on conversion accuracy and conversion speed of ADC of RFFE.The dynamic range of system could be deteriorated with the increase of bandwidth which introduces larger interference.While receiving signals,the most common issue faced is that small signals are too weak relative to non-linear distortion products of large signal or strong interference,which causes the decline of testing and receiving of receivers.The issue reacts on SFDR of RFFE directly for enhancing SFDR contributing to testing small signals effectively.Through theoretical derivation and computational analysis,to explore the source of the noise and nonlinear distortion in digital receiving front-end.Then analyzed the source of non-linear distortion whose result could indicate that the mean factor of testing small signal is non-linear distortion.Only does non-linear distortion solve,SFDR could be improved.Comparing Volterra series model,polynomial model and blocking model,studying on accuracy and algorithm complexity,it points on superiority on relevance grade and fitting precision of Volterra series model.But due to the linear and nonlinear distortion transfer function change in work within the bandwidth is relatively complex,on the series will set higher order number and the memory depth,to ensure the accuracy of fitting and complexity will increase accordingly.To get the compensation simulation of non-linear distortion with the means called classical Wiener Filtering Method,and results show that the improvement of the system comes to 8dB.To study on distortion compensation based on MMSE and KDMS,the former use of instantaneous amplitude of the input signal,while the latter using the statistic information.The performance of MMSE is more superior in increased with the increase of the depth of memory,there were 16 dB in memory depth of 6.In a system whose nonlinear distortion is not very serious,linear memory distortion error influence is also cannot beignored,the simulation results show that the system has 22 dB linearity improvement based on KDMS.Two algorithms are better than wiener filtering with varying degrees of improvement.