多径衰落环境中具有调制识别能力的盲均衡新算法

New Blind Equalization Algorithm with Ability of Modulation Classification Under the Condition of Multipath Propagation

常数模盲均衡算法在对非常模信号进行均衡时稳态均方误差不能收敛至零,且不具备信号调制识别能力.因此,依照常数模算法代价函数的构造方法定义了一种余弦代价函数,提出了基于该代价函数的盲均衡新算法.该余弦代价函数可将BPSK、M-PAM、M-QAM信号的不同星座点映射至原点,从而使新算法在对上述常模或非常模信号进行均衡时稳态均方误差均能收敛至零,更重要的是新算法能够在多径衰落环境下实现对上述信号的调制识别.理论分析和仿真结果证明了新算法的优良性能.

For nonconstant modulus signals,the steady-state mean square error（MSE）of constant modulus algorithm（CMA）for blind equalization can not converge to zero and CMA can not achieve modulation classification.According to the construction method for CMA′s cost function,a cosine cost function is defined and a new blind equalization algorithm based on it is proposed.Such cost function can take all the constellation points of BPSK,M-PAM and M-QAM signals mapped to the origin of constellation diagram,so the steady-state MSE of the proposed algorithm can converge to zero for both of constant modulus and nonconstant modulus signals.Furthermore,for above signals the proposed algorithm has the modulation classification ability under the condition of multipath propagation.Theoretic analysis and computer simulation results demonstrate that the proposed algorithm offers high performance.