Research On Background Noise Removing In Electromagnetic Radiation Signals Based On Fast ICA

In recent years, with the rapid development and the development electronic communication technology more and more electronic, electrical and communications equipment comes into people’s daily life. However, the effects of the electromagnetic radiation produced by these devices is being noticed. Meanwhile, the sensitivity of many electronic devices has become increasingly high, the frequency and intensity of the around electromagnetic radiation are also getting higher, so the EMC (Electro Magnetic Compatibility) discipline also becomes increasingly important. In this discipline, the electromagnetic radiation measurement technology is very important, however, the actual measurement, the co-frequency ambient noises in the actual environment result in inaccurate measurement of electromagnetic radiation. To solve this problem, to achieve the accurate measurement of electromagnetic radiation, and get the correct parameters for the electromagnetic environment evaluation. It is particularly important to research on the co-frequency ambient noises removing.Firstly, this paper research on the basic characteristics of the electromagnetic radiation signal, the basic concept of the corona discharge, the physical process and the factor that affect the conductor surface electric field. According to the literature method of calculating the electromagnetic radiation signal on the high-voltage transmission lines, the co-frequency ambient noises is studied. A comparative analysis of signal characteristics described in the literature is done, the electromagnetic radiation model is established and the accurate electromagnetic radiation signal is got.Secondly, the theory and application of blind signal separation algorithm is studied, that is the extraction of the electromagnetic radiation signal from mixed the co-frequency ambient noises in the actual measurement process. The co-frequency ambient noises is removed from the measured electromagnetic radiation signal, the fast independent component is used in theory analysis and simulation of mixing and separation. The simulation results show that, using the fast independent component analysis (fastICA) algorithm can achieve the separation of electromagnetic radiation signal, signal similarity coefficient is0.9688.Finally, the similarity coefficient of electromagnetic radiation signal between the source signal and after the separation, separation effect and crosstalk error are used for evaluation the performance of the fastICA algorithm, and the fastICA algorithm and adaptive interference cancellation algorithm are compared.The parameter for the accurate measurement of electromagnetic radiation and electromagnetic environment assessment is particularly important. The removing of the co-frequency ambient noises is very significant to the assessment of electromagnetic environment.