Abstract: The problems of multipath attenuation and high background noise in the channel of Low voltage Power Line Communication（L-PLC） system can not be ignored. Discontinuous transmission of L-PLC based on Fast Independent Component Analysis（FastICA） is proposed to improve the anti-jamming ability of system. On the basis of the establishment of noise model of the low voltage power line channel, this paper uses the FastICA to denoise the voice signal having strong noises on the low voltage power line, does voice activity detection for the processed voice signal, and eventually achieves the speech signal which meets certain quality requirements. The Matlab simulation experimental results show that under low SNR, the described method successfully achieves the goal of conditional continuous speech signal transmission, and improves to some extent the quality of the voice transmission.

Key words: Fast Independent Component Analysis（ICA）, Low voltage Power Line Communication（L-PLC）, noise model, discontinuous transmission, voice activity detection

摘要： 在低压电力线通信（L-PLC）中，多径衰减和信道背景噪声大的问题不可忽视，为此提出了一种基于快速独立分量分析（FastICA）的L-PLC间断传输提升抗干扰能力的方法。在建立L-PLC信道噪声模型的基础上，利用FastICA对强噪声干扰的语音信号进行去噪处理，并对处理后的信号进行话音活性检测，最终获得满足质量要求的语音段。Matlab仿真实验结果表明，在低信噪比环境下，该方法成功实现了有条件的语音信号间断传输，并在一定程度上提高了语音传输质量。

关键词: 快速独立分量分析, 低压电力线通信, 噪声模型, 间断传输, 话音活性检测