Study On ICI Cancellation Algorithms For OFDM Systems

Orthogonal frequency division multiplexing (OFDM) has become a key technique in broadband wireless access and the next generation mobile communication system. Extensive attention has been paid to OFDM due to its high spectral efficiency, high data rate and well anti-multipath ability. As a multi-carrier modulation technique one major disadvantage of OFDM is it is sensitive to the frequency offset. Frequency offset can cause the loss of orthogonality among different subcarriers, then the Inter-Carrier Interference (ICI) occurs.This thesis mainly analyzes the ICI problem in OFDM systems caused by time- variant multipath channels, studies several ICI cancellation algorithms based on simulation and performance evaluation. In this thesis firstly the cause of ICI and the affect of ICI are analyzed, then three main methods to suppress ICI, including Self-Cancellation(SC), Maximum Likelihood Estimation(MLE), and Segment equalization are introduced. then the performance of three methods are compared based on MATLAB simulation, the implement complexity of these methods are also analyzed and compared. Based on the BER curve the disadvantages and advantages of these methods are also compared. Based on above study, two improved schemes are researched in this thesis. The first scheme is an improved segment equalization scheme based on segment window filter. By choosing a specific data window with sidelobe characteristics, the pervasion of channel impulse response after inverse transform can be reduced. With this scheme we can decrease estimation error and improve the capacity for the elimination of ICI; in the second scheme, a new method which unite Self-Cancellation scheme and Improved-MLE scheme is investigated. Though its complexity is increased to some extent ,it has the same band efficiency with the scheme of MLE and has better performance than MLE. The new united scheme can also improve the capacity for the elimination of ICI.