Abstract: This paper proposes a novel Varying Step Size Modified Constant Modulus Algorithm（VSS-MCMA）. And on this basis, this paper presents another new decision directed variable step size MCMA（DD-VSS-MCMA） based on the Mean Square Error（MSE）, that is, the VSS-MCMA is switched to the DD algorithm when the MSE is small enough. Furthermore, the constellation matched error（CME） function is inserted into the cost function of the DD-VSS-MCMA and the CME-DD-VSS-MCMA algorithm is put forward. The simulation results show that the proposed VSS-MCMA can simultaneously improve the convergence and the steady-state residual ISI performance compared with MCMA, DD-VSS-MCMA can further reduce the residual Inter-Symbol Interference（ISI）, and the presented CME-DD-VSS-MCMA can speed up the convergence and can attain slower residual ISI.  

Key words: modified constant modulus algorithm, variable-step-size, decision-directed, constellation matched error, residual Inter-Symbol Interference（ISI）

摘要： 以修正常数模算法（MCMA）为基础，用变步长代替固定步长，提出一种新的变步长修正常数模算法（VSS-MCMA），并在此基础上提出一种基于均方误差判决引导（DD）的双模式盲均衡算法（DD-VSS-MCMA），该算法在收敛阶段采用收敛性能良好的VSS-MCMA算法，当剩余均方误差达到足够低时切换到DD算法，又在此算法的代价函数中加入星座图匹配误差函数（CME），进一步提出CME-DD-VSS-MCMA算法。计算机仿真结果表明：与MCMA算法相比，提出的VSS-MCMA可以同时提高收敛速度和降低稳态剩余码间干扰（ISI），DD-VSS-MCMA进一步降低了稳态剩余ISI，CME-DD-VSS-MCMA显著地加快了收敛速度并获得了更好的稳定性能。

关键词: 修正常数模算法, 变步长, 判决引导算法, 星座图匹配误差函数, 剩余码间干扰