Research On The Adaptive Compensation Methods Of The Phase Noise In The OFDM System

Nowadays, the high speed broadband wireless communication, especially the 4th generation mobile communication technology (4G) is becoming one of the hot topics in the field of communication. The most important characters of the 4G communication are the greater frequency efficiency. Moreover, it can confront the frequency selective fading and the multi-path fading. How to realize the high speed digital communication in limit frequency band or under the condition of poor channel is the primary topic in the field of wireless communication.Among the wireless communication systems at present, the Orthogonal Frequency Division Multiplexing (OFDM) system has been widely used in the Digital Audio Broadcasting (DAB) system, the Digital Video Broadcasting-Terrestrial (DVB-T) system, wireless LAN and the alternative standard of 4G communication system. The widely using of OFDM system is due to that it can confront the distortion of the dispersive channel effectively, and it has a higher frequency utilization ratio. Moreover, the OFDM system can confront the pulse noise effectively and has lower system complexity.However, the OFDM system is sensitive to the linear phase noise due to frequency shift and the Wiener phase noise due to the nonlinear character of the Voltage Controlled Oscillator (VCO). These kinds of two phase noises can bring in the serious Inter-Carrier Interference (ICI) and the Common Phase Error (CPE) to the OFDM system, which can affect the performance of the OFDM system seriously.Under this background, an adaptive algorithm, which has of low complexity, better compensation performance and is easy to realize on the hardware platform, the Least Mean Square (LMS) algorithm is utilized to compensate the linear phase noise and the Wiener phase noise. The research of the compensation method includes the theoretical analysis, the performance simulation of the algorithm and the hardware platform design.Firstly, as the basic theory, the producing reasons and the characters of these two phase noises are listed and analyzed. Through analyzing of the characters of these two phase noises and comparing with the compensation characters of the adaptive filter algorithm, the adaptive algorithm is selected as the foundation of the compensation algorithm. Furthermore, the basic theory and structure of the adaptive algorithm are introduced. The characters of compensation objectives of the adaptive algorithm are that the expected signal can be formed, the variance of the target signal should be small and has converged error signal. The LMS algorithm is selected as the foundation of the compensation algorithm among all the adaptive algorithms, because it is easy to realize, fast to converge, better to compensate.Secondly, the effect of the linear phase noise on the OFDM system is analyzed. According to the character of the linear phase noise, a compensation method named UE-LMS algorithm of LMS algorithm combined with the phase offset estimation is adopted. The compensation process is to pre-estimate the phase offset before DFT transform and to compensation the remnant phase offset of the serial signals after DFT transform by the LMS algorithm. The feasibility of this compensation method is deduced theoretically. And the realizing parameters, compensation ability and the converge speed are analyzed by the Matlab simulation tool. Through the theoretical analyzing and the simulation analyzing, it can be found that the compensation ability of the UE-LMS method is good and has fast converge speed and short Pilot Symbols (PS).Thirdly, the effect of the Wiener phase noise on the OFDM system is analyzed. According to the character of the Wiener phase noise, a compensation method named AO-LMS algorithm of LMS algorithm whose objective is the phase angle offset is adopted. A classical method using the adaptive algorithm to compensate the Wiener phase noise named the Finite Impulse Response-Adaptive Linear Equalizer (FIR-ALE) method is discussed and improved in this article. Furthermore, the realizing parameters, compensation ability and the converge speed are analyzed by the Matlab simulation tool. Through the theoretical analyzing and the simulation analyzing, it can be found that the compensation ability of the AO-LMS method is better than that of the FIR-ALE method.Fourthly, the feasibility of the combined compensation method for both the linear phase noise and the Wiener phase noise using the LMS algorithm is discussed. An adaptive compensation system is presented, which is to pre-estimate and treat the phase offset of the linear phase noise at first and then compensate the remnant phase offset and the Wiener phase noise. The realization method of this combined compensation system is analyzed theoretically. The compensation ability of this combined system is analyzed by Matlab. The simulation results indicate that the compensation ability is good and can combine the UE-LMS method and the AO-LMS method well.Lastly, the Field Programmable Gate Array (FPGA) is adapted to implement the hardware design of the UE-LMS method at the receiver of the OFDM system. The difficulties to designing the platform of the OFDM system and the LMS algorithm are analyzed. A simulation analyzing method by the combination of Quartus waveform and Matlab simulation is presented. By complexity comparison between LMS algorithm and another adaptive algorithm– the Recursive Least Squares (RLS) algorithm, the significance of selecting of LMS algorithm is proved again. By the designing of the hardware, the feasibility of the LMS algorithm using in the practical OFDM system is verified and the application value of the UE-LMS algorithm is proved also.