正交频分复用(OFDM)技术由于具有频谱利用率高、能够有效对抗频率选择性衰落等优点，受到越来越多人们的关注，成为后三代移动通信的关键技术。但是由于无线信道的多径性和时变性，造成了接收端OFDM符号子载波间的干扰，为了能够准确地解调OFDM信号，就必须对接收到的信号进行信道估计。 本文分析了导频分布的两种方案：1)块状导频分布，对基于块状导频分布方案的慢衰落信道估计算法：最小平方(LS)和线性最小均方误差(LMMSE)算法进行了计算机仿真，比较了二者的估计性能，以及各自的优缺点；2)梳状导频分布，对基于这种分布方案的快衰落信道估计算法进行了研究，对两类经典算法：线性插值法和基于变换域的插值法分别进行了理论分析和计算机仿真。 本文的主要工作是提出了三种OFDM系统的信道估计改进算法：1)使用了新的插值方法的改进算法：针对线性插值算法估计误差较大，而基于变换域的插值法计算量太大的特点，利用三次样条插值算法和最小均方误差(MMSE)方法代替线性插值和基于变换域的插值算法对信道进行估计；2)基于LMMSE的改进算法：LS在实现过程中未考虑噪声和子载波干扰(ICI)的影响，影响了系统的传输性能，本文提出了用LMMSE代替LS算法对导频处的信道函数进行估计；3)导频非均匀分布的改进算法：导频均匀分布时，由于无线信道的特性，造成了边缘数据符号的误码率远远大于中央的误码率，从而总的传输误码率也随之增大，本文利用了牛顿插值法，当边缘和中央的估计误差相当时，计算出二者不同的导频分布宽度，再进行信道估计的过程。并且通过计算机仿真证明了三种改进算法与原有算法相比，在高速传输时候的优越性，表明了其更适合于第四代移动通信的高速要求。

OFDM(orthogonal frequency division modulation) has gained more and more considerable attraction, and it is regarded as the key technology for the next mobile communication system, due to its high bandwidth efficiency and robustness against frequency selective fading channel. The radio channel is time-variant, frequency-selective and noisy, so the received OFDM signals are corrupted by intercarrier interference. To demodulate the OFDM signals accurately, channel estimation algorithms are adopted. This paper examines the slow-fading estimation algorithms based on block pilot arrangement: LS and LMMSE algorithms, as well as the fast-fading channel estimation algorithms based on comb-type pilot arrangement: linear and transform domain interpolation algorithms. The performance of the various algorithms are analyzed by matlab simulation. The paper gives three advised algorithms: 1) The linear interpolation and transform domain interpolation algorithms are replaced by the spline cubic and MMSE algorithms to reduce the complexity of achieving the transfer function; 2) LMMSE algorithm is presented to estimate the transfer function of the pilot signals;3) To balance the bit error rate of the data signals in the center and brim, the paper uses the asymmetric pilot arrangement. The values of the distance of the center and the brim are found by Newton interpolation. The simulation results demonstrate the three advised algorithms are preferable choices for the OFDM system.