凸组合实时判决反馈稀疏水声信道盲均衡算法

针对高速水声通信信道稀疏特性,提出了一种凸组合实时判决反馈盲均衡算法。将盲均衡器分为保持均衡器支路和稀疏均衡器支路,以保持均衡器能量和权系数的瞬时梯度为判据,对稀疏均衡器支路对应抽头进行实时稀疏化处理。算法中避免设置稀疏化阈值,对不同稀疏水声信道和通信信号具有通用性,且对于时变稀疏水声信道可以利用保持均衡器支路恢复稀疏均衡器支路置零抽头系数,使算法对信道具有较强跟踪和冷启动能力。典型稀疏水声信道条件下的仿真结果证明,凸组合实时判决反馈稀疏水声信道盲均衡算法性能稳健,与全阶判决反馈盲均衡算法相比,计算简单,收敛速度快,稳态剩余误差小,有利于算法在水声通信系统中的推广应用。

Convex Combination Real Time Decision Feedback Sparse Underwater Acoustic Channel Blind Equalization Algorithm

According to the sparse characteristics of underwater acoustic channel in high speed communication system, a convex combination real time decision feedback blind equalization algorithm was proposed. The blind equalizer is divided into two branches, called maintaining equalizer branch and sparse equalizer branch. And the sparse equalizer branch carries sparse processing according to the instantaneous gradient and the energy of the maintaining equalizer branch. The proposed algorithm avoids sparseness threshold setting, which can be applied to different sparse underwater acoustic channel and communication signal conditions. Furthermore, the maintaining equalizer branch can recover the sparse equalizer branch in time for the time variant sparse underwater acoustic channel, which makes the algorithm have strong tracking and cold starting ability. The simulation results of typical sparse underwater acoustic channel show that the convex combination real time decision feedback sparse underwater acoustic channel blind equalization algorithm has some advantages over full order decision feedback blind equalization algorithm, such as robustness, simple calculation, faster convergence rate and lower steady state residual error, which is beneficial to the application of the algorithm in underwater acoustic communication system.