基于卡尔曼滤波提出了两种相干光正交频分复用（CO-OFDM）系统的相位噪声补偿算法，这两种算法在发射端的时域均插入导频，并在接收端对导频进行卡尔曼滤波，最后利用插值算法补全全部子载波的相位噪声。仿真结果表明，基于最小均方误差（MMSE）准则的判决反馈算法在相位噪声比率为10-4时，系统误码率约为10-4，并且出现了错误平层，而基于卡尔曼滤波所提出的两种相位噪声算法在大相位噪声的情况下仍然具有较好性能且能有效地降低错误平层，因而所提出的相位噪声补偿算法能改善CO-OFDM系统的性能。

The novel two phase noise compensation algorithms based on Kalman filter in coherent optical orthogonal frequency division multiplexing(CO-OFDM) systems are proposed. The two algorithms insert the pilot frequency into the time domain of the transmitter and perform the Kalman filter of the pilot at the receiving terminal, and finally apply the interpolation algorithm to compensate the phase noise of all the OFDM subcarriers. The simulation results show that the two proposed algorithms based on Kalman filter have better performance of the bit error ratio(BER) and can effectively reduce the error floor at the condition of the large phase noise while the BER of the decision feedback algorithm based on the minimum mean square error (MMSE) criterion tends to 10-4 and the error floor occurs at phase noise rate of 10-4. Therefore, the proposed phase noise compensation algorithms can improve the performance of CO-OFDM communication systems.