Research And Simulation On Inter-carrier Interference Mitigation In OFDM System

In recent years, Orthogonal Frequency Division Multiplexing (OFDM) is widelyused as its robustness against multipath fading, high spectrum efficiency, andconvenience of combining with multiple input multiple output (MIMO). It has beenapplied in wireless communication systems such as Digital Audio Broadcasting (DAB),Digital Video Broadcasting (DVB) and so on. The3rd Generation Partnership Project(3GPP) has selected OFDM as the key downlink technology in Long Term Evolution(LTE).As OFDM utilizes multiple subcarriers to transmit signal, it is easily to be affectedby carrier frequency offset (CFO). CFO can destroy the orthogonality amongsubcarriers and results in severe inter-carrier interference (ICI). The ICI may maskdesired signal and make signal detection difficult. CFO mainly comes from Dopplerspread and oscillator mismatch between the transmitter and receiver. High Dopplerspread will make channel become time selective, and with the inherent multipath fadingin channel, which generate the doubly (frequency and time) selective channel. Since ICIcan decrease the detection performance, it is necessary to propose methods that canmitigate ICI.The thesis will only discuss ICI from Doppler spread, the structure shows asfollows:Chapter1describes the development of wireless communication, introduces thefoundation of multicarrier system, shows the structure of OFDM and MIMO, andexplains the main work of the thesis.Chapter2shows the OFDM system in high speed environment, exploits someformulas to describe the relationship between transmitting and received signal, andsimply analyses the cause of ICI. At last, it summarises classical methods to detectsignal in high speed environment.Chapter3introduces a banding approximating model in channel coupling matrix that can lead to some low complexity detection methods in doubly selective channel,such as the parallel iteration cancellation and MMSE-SD，we also propose someimprovement about these two methods. In addition, we introduce a method that utilizesICI depression matrix which can make the SINR maximal, which leads the equivalentcoupling matrix verge to a diagonal matrix. Matlab simulations show the effect of thesemethods.Chapter4discusses the case in multi-antenna system which contains MIMOextended detection, which also utilizes the banding approximating model. And it showsthe similar method that using a ICI depression matrix which makes the equivalent SINRmaximal. With the different definition of SINR, it leads to different equalizationapproaches at receiver. At last, we propose a new system model named partial doublyselective channel. That means only part channels between transmitters and receivers aredoubly selective, the rest are still time invariant. And we propose a low complexitydetection method.Chapter5summarizes the whole dissertation, and puts forward some suggestionsabout further research in ICI suppression under high speed environment.