为了提高噪声消除的鲁棒性,传统的通用旁瓣消除器采用固定波束形成作为输出的系数约束自适应滤波器,采用CCAF的输出作为输入的标准约束自适应滤波器。此系统的跟踪性能随信号频率变化而改变,并且由于波束带宽的影响使得目标语音在低频发生畸变。采用优化波束形成方法来得到较宽的带宽,对输出进行后滤波处理来消除残余噪声。试验表明本方法可明显提高语音信噪比。
To improve the robustness of noise cancellation, the conventional generalized sidelobe canceller (GSC) consists of coefficientconstrained adaptive filters (CCAF) which take the fixed beamforming as input. The tracking range varies with the signal frequency, and the low frequency target is distorted due to the beam bandwidth. A new structure is proposed in which the fixed beamformer is replaced by the optimized beamformer to achieve broad bandwidth ,and the output is filtered by a post-filtering to eliminate the residual noise. Simulation results show that the SNR is increased significantly.