为进一步提高盲均衡器的收敛速度，减少稳态误差，提出针对两个不同p和不同步长的CMA盲均衡滤波器的凸组合方案，以适用于服从alpha稳定分布的脉冲噪声环境。利用分数低阶统计量、能量守恒以及分离假设等条件，推导出该方案的超量均方误差(excess mean square error, EMSE)理论公式。仿真结果表明,该方案能够保留各分量盲均衡滤波器的优势，可以在提高系统收敛速度的同时减少稳态误差，进而验证了该方案的准确性与合理性。

In order to further improve the convergence speed of the blind equalizer and reduce the steady-state error, a convex combination scheme for two constant modulus algorithm (CMA) blind equalization filters with different p and asynchronous length was proposed to be suitable for impulse noise environment subject to alpha stable distribution. The theoretical formula of excess mean square error (EMSE) of the scheme was derived by using fractional lower order statistics, energy conservation and separation hypothesis. Simulation results show that the scheme can retain the advantages of each component blind equalization filter, improve the convergence speed of the system to reduce the steadystate error, which verify the accuracy and rationality of the scheme.