中文摘要：

传统的盲均衡算法需要预先具有均衡器阶数的先验知识，提出一种带阶教检测的信道盲均衡算法，采用进化算法结合常数模代价函数对均衡器阶数、均衡器权系数进行联合进化操作寻优，同时实现阶数、权系数估计。利用模拟浅海水声信道数据进行的仿真实验结果表明：本算法可有效避免常数模算法在低估阶数时造成的性能下降，适用于未知阶数情况下的信道盲均衡。

英文摘要：

A common assumption in the conventional blind equalization methods is that the order of equalizer is known in advance. However, such information is in fact not available. A joint blind order-detection and equalization algorithm was proposed to avoid the dependence on order information. The problem was tackled with an evolutionary programming strategy, in which the order, as well as the taps of the equalizer were evolved simultaneously with the objective function of constant modulus algorithm （CMA） adopted as fitness function. The simulation results on shallow water acoustic channel demonstrate that the proposed algorithm achieves satisfactory performance and robustness upon order information.