Study On The ICI Mitigation Technique In Mobile OFDM Systems

With the advantages of robustness against muti-path fading channel, high spectrumefficiency, simple channel equalization and easy hardware implementation, orthogonalfrequency division multiplexing (OFDM) is the key physical layer technology inbeyond3rd generation communication systems. Because the inter-carrier interface (ICI),which caused by doppler spread in mobile environment, destroys the orthogonalityamong the subcarriers and degrades the syetem preformance, ICI mitigation is one ofthe pivotal technique in OFDM mobile communication.This paper studys on ICI mitigation and FPGA implementation depending onpractical OFDM system programs. The main contents of the paper are listed as follows:①Mainly analyze the causes and effects of ICI in mobile OFDM environment.Study and simulate the ICI effect of single doppler frequency offset and doppler spread.Simulation results show that doppler spread is the major reason that responsable fortime-varying characteristics of mobile channel.②ICI mitigation algorithm needs to comprehensively consider the mitigationeffect, spectrum efficiency and implementation complexity. Self-Cancellation algorithmsuppress ICI effects by means of changing system parameters design. Validate ICImitigation effect of coding and time-domain windowing and compare performance oftwo algorithms by simulation, then discover that it has the shortcoming of low spectrumutilization rate together with error floor. Channel estimation algorithms, which is basedon frequency-domain equalization, employ the channel response matrix estimated bypilot assisted criterion to suppress ICI. Analyze, simulate and evaluate the performanceof traditional LS and LMMSE algorithms, in addition, compare the performance ofcommon piolt arrangements in different channels. Simulation results show thattraditional channel estimation algorithms have the performance limitation.③We design and study the simplified G matrix algorithm to overcome theperformance limitation of traditional one-dimensional channel estimation algorithm inmobile environment. This method, with better performance than traditional channelestimation algorithms and low implementation complexity, takes into account theimpact of adjacent subcarriers and linearly approximates the time-varying channel bycalculating the slope of time-domain impulse response. In order to reduce the errorcaused by directly using IFFT to calculate time domain pluse response in simplified G matrix algorithm, a modified time frequency transform method is proposed. Simulationresults show that modified method has better performance on MSE and BER.④Complete the FPGA implementation of simplified G matrix algorithm, slovethe issues that ICI degrades the performance in practical communication systems. Toachieve less hardware resources, some optimization programs such as the time divisionmultiplexing of logic and memory resources are proposed. The test results show that theprogram designed in this paper functions properly, consume fewer resources andsuppress ICI effectively.