An Impulse Noise Suppression Algorithm Based on Dual Interface Diversity and Data Samples

Hybrid power line and wireless communication has wide application prospect in the Internet of Things. How to suppress the pulse noise of the power line channel in dual interface communication and improve the reliability of communication system is one of the key problems to be solved urgently. Aiming at the interference problem of power line burst pulse noise (BIN) in the power line and wireless dual interface communication, this paper put forward a BIN suppression algorithm based on diversity signal cancellation and adaptive threshold estimation (DSC-ATE), by using the independence and difference of power line and wireless channel and the consistency of diversity signal. In DSC-ATE, the wireless channel is used to assist the power line channel to extract the noise samples, and then the pulse noise threshold is estimated adaptively through the noise samples, finally the pulse noise is separated from the received signal. The traditional OFDM system has the problem of peak-to-flat ratio PAPR, which is difficult to obtain the position information of pulse noise directly in the power line channel. Based on the dual-interface communication architecture, a noise sample extraction method based on diversity signal cancellation is proposed, which improves the accuracy of noise detection and can be applied to the optimal threshold prediction. The simulation results show that after diversity signal cancellation, the effect of PAPR is minimal because only background interference remains, and the precise position information of impulse noise can be obtained by using nonlinear transformation, which provides data support for the optimal threshold estimation algorithm. Secondly, the sample space of impulse noise is constructed by means of signal cancellation processing in diversity transmission, in which the noise data of low SNR ensures the diversity of samples. Then, an optimal threshold estimation algorithm based on noise sample data and nonlinear functions is proposed, which can minimize the bit error rate of the communication system and does not require the prior information of impulse noise. The simulation result shows that the performance of threshold precision, transmission rate, and bit error rate under different SNRs of the algorithm is superior to the traditional algorithms such as the weighted combination criterion (WCC) and siegert criterion (SC). The proposed algorithm also has certain robustness and robustness under different channel fading and noise parameters. Finally, aiming at the problem of pulse noise elimination in non-stationary environment, the noise threshold is iteratively estimated by using the data sample and the objective function minimization/maximization algorithm, so as to determine the location information of the pulse noise. The convergence rate of the algorithm is adjusted by importing parameters such as learning rate and discount factor, and the threshold can be dynamically adjusted according to the changes of the environment, so as to achieve the effective compromise between the robustness and convergence rate of the algorithm. The simulation result shows that when the learning rate of the algorithm is fixed, the larger the data sample length is, the smaller and more stable the threshold fluctuation is. However, the large data sample length will increase the computational burden, storage burden, and data sample acquisition time, which slow down the update speed of the threshold. In the future, the proposed algorithm DSC-ATE can be combined with sparse theory and compressed sensing for other noise models such as narrowband noise, to further improve the universality of the algorithm. © 2023 Chinese Machine Press. All rights reserved.