Wavelet transform based feature extraction of signal for zero-crossing modulation technique

The zero-crossing modulation technique(ZCMT) is commonly used for control, automatic remote meter reading and monitoring on power distributed network. But the extracted raw signal contains a lot of interferential noises in ZCMT. The problem is to capture the true signal. This paper describes wavelet transform algorithm based on wavelet transforms theory, proposes the principle of de-noising signal. In experiment, a second-order of Daubechies wavelet is chosen as wavelet function, soft thresholding function and fixed threshold value are adopted for canceling signal noise. The result indicates that noise can be eliminated and signal features are extracted by using wavelet Transform in the zero-crossing modulation technique. The power line communication is a branch of communication technology. The low voltage power line is an ideal communications medium for digital communication[]]. Because the infrastructure of low voltage power lines is already present and the physical link is reliable, the main advantage of power line communication is that initial cost is relatively low. It includes the power line carrier (PLC) communication and the zero-crossing modulation technique(ZCMT). The carrier approach adopts frequency modulation or phase modulation method to carry out digital communication by using high voltage power line with 10kv or higher than 10Kv to transmit signal, because the interference on high voltage power line is little. It is a challenge to carry out carrier communication because there are a lot of interferences over low voltage power line[2]. ZCMT is commonly used for control, automatic remote meter reading, demand-side management and monitoring on power distributed network. Its communication distance is longer than PLC But in ZCMT the extracted raw signal contains a lot of interferential noises, wavelet transform is employed to detect the raw signal, the noises are canceled and the characteristic of true signal is extracted.