| OFDM (Orthogonal Frequency Division Modulation) has gained more and more considerable attention, and it is regarded as the key technology for the next mobile communication system, due to its high bandwidth efficiency using rate and robustness against frequency selective fading channel. These years, it has being used in DVB, DAB, WLAN and so on. The radio channel is time-variant, frequency-selective and noisy, so the received OFDM signals are corrupted by inter-carrier interference. To demodulate the OFDM signals accurately, channel estimation algorithms are adopted. In this thesis, we are working on the topic based on some of the results of the precedents.At first, introducing the principle and features of OFDM system, analyzing the superiority and deficiency of OFDM, and researching the channel estimation for OFDM system, based on the channel modeling and the preferences. Tow kinds of channel estimation algorithms are introduced: based on block-type pilots and based on comb-type pilots. The channel estimation algorithms based on block pilot includes: LS, MMSE and LMMSE algorithms, as well as the channel estimation algorithms based on comb-type pilot includes: linear and transform domain interpolation algorithms. The superiority and deficiency of the various algorithms are analyzed by Matlab simulation.Two advised algorithms are given: (1) LMMSE algorithm with SVD decompound is presented to estimate the transfer function of the pilot signals, insteading of MMSE algorithm;(2) The linear interpolation and transform domain interpolation algorithms are replaced by the spline cubic and MMSE algorithms based on Winner filter to reduce the complexity of achieving the transfer function. Because of the better capability and the less calculation, outcome of the simulation shows that the advised algorithms are preferable choices for the OFDM system anti even more fit for the next mobile communication system. |
