Low-voltage Power Line Signal De-noising Based On Independent Component Analysis

High-speed power line communication (PLC) refers to the use of low-voltage power line as a transmission medium to transport high-speed data and voice services. In recent years, with the rapid development of Internet technology, using power line for information communication has become a hot research field now. However, the low-voltage power network is connected with thousands of households, a wide range of electrical equipment will produce complex noises. As a result,its characteristics are not ideal while the power lines as a communication channel. At present the main methods of power line signal denoising are Orthogonal Frequency Division Multiplexing (OFDM) technology, spread spectrum technology, wavelet transform, and so on. However, when use of OFDM technology, in the transmission process, the complex characteristics of the noise in the power lines will interfere with the carrier signal in the channel transmission, makes the quality of signal transmission decline in varying degrees. Spread spectrum communication can make the lower error rate, but it sacrificed much of the space transmission bandwidth. Wavelet technology for denoising improves the calculation and spends a long times on filtering. Therefore, using an effective method to remove the low-voltage power line noise to improve the quality of transmission and protect the reliability of the power line communication is particularly important.In this paper, a denoising method for low voltage power line signals based on independent component analysis (ICA) is proposed. And we introduce low-voltage power line channel noise characteristics and principles of independent component analysis in detail. In this paper, channel noises of low voltage power line and the principle of ICA are introduced. The proposed method is applied to denoising low voltage power line signals using FASTICA Algorithm based on negentropy and it is compared with another ICA method. The experimental results show that the denoising performance of the proposed method is easier to use and possesses a good denoising effect and good adaptive ability.