水声通信中稀疏信道均衡算法优化

Optimization of Sparse Channel Equalization Algorithm in Underwater Acoustic Communication

为应对复杂的水声信道环境,提高信道均衡算法的收敛速度和误码率性能,文中提出了一种零值吸引稀疏控制成比例最小误码率判决反馈均衡算法。该算法在稀疏控制成比例最小误码率判决反馈均衡算法的基础上,通过在目标函数中加入近似l0范数的稀疏约束,迫使均衡算法在迭代过程中将幅值小的均衡器抽头向零值吸引,以加快均衡算法的初始收敛速度;同时在信道均衡过程中引入锁相环技术,以消除抖动相位噪声带来的影响。传统的锁相环技术都是基于最小均方误差准则的,但现有文献和相关实验仿真已经证明,当系统的均方误差最小时,误码率不一定最小。针对此问题,文中提出了一种基于最小误码率准则的锁相环相位追踪算法,并将其嵌入稀疏均衡算法中。在Matlab平台上,分别在实际采集的静态水声信道和实际时变水声信道条件下进行了实验,结果表明:加入近似l0范数约束的稀疏控制成比例最小误码率判决反馈均衡算法,在没有时变相位噪声影响下的收敛速度更快;在有时变相位噪声影响的信道条件下,基于最小误码率准则的锁相环相位追踪算法相较于基于最小均方误差准则的锁相环相位追踪算法,收敛速度更快,误码率性能更优。

In order to cope with the complex underwater acoustic channel environment and improve the convergence speed and symbol error rate performance of the channel equalization algorithm,this paper proposed a zero attraction sparse control proportional minimum symbol error rate decision feedback equalization algorithm.On the basis of the proposed sparse control proportional minimum symbol error rate decision feedback equalization algorithm,this algorithm added a sparse constraint of approximate l0norm to the objective function,which pulls small amplitude equalizer taps toward zero.At the same time,phase-locked loop technology was introduced in the channel equalization process to eliminate the influence of jitter phase noise.The traditional phase-locked loop technology is based on the minimum mean square error criterion.However,existing literature and related experimental simulations have demonstrated that,when the mean square error of the system is the smallest,the symbol error rate is not necessarily the smallest.Aiming at this problem,a phase-locked loop phase tracking algorithm based on the minimum symbol error rate criterion was proposed and embedded in the sparse equalization algorithm.On the Matlab platform,experiments were carried out on the static underwater acoustic channel and the real time-varying underwater acoustic channel,respectively.The results show that the sparse control proportional minimum bit error rate decision feedback equalization algorithm with approximate l0norm constraint converges faster without the influence of time-varying phase noise;under the channel condition affected by time-varying phase noise,the phase tracking algorithm based on the minimum bit error rate criterion has faster convergence speed and better bit error rate performance than the fragrance tracking algorithm based on the minimum mean square error criterion.