A Broadband Spectrum Signal-to-Noise Separation Method Based on Neighbor Statistics

Abstract

The present invention relates to a broadband spectrum signal-to-noise separation method based on neighbor value statistics, which includes the following steps: Step S1, providing a broadband spectrum in a given radio wave environment; Step S2, selecting a spectrum sample; Step S3, calculating the standard deviation of the spectrum noise; Step S4, determine the discriminant value for the comparison of adjacent values; Step S5, calculate and obtain the initial noise signal used for the comparison of adjacent values; Step S6, obtain spectral noise data through the spectral noise extraction method; and Step S7, perform signal extraction to realize signal-to-noise separate. The invention is simple and practical, has fast calculation speed and relatively high accuracy, and is suitable for real-time frequency spectrum analysis and signal statistics.

Description

technical field [0001] The invention relates to the technical field of radio astronomy, in particular to a broadband spectrum signal-to-noise separation method based on neighbor value statistics. Background technique [0002] Large-aperture radio telescopes have extremely high system sensitivity, and there are many electronic devices in the system, between systems and within the site. With the development and application of high-frequency electronic technology and high-speed digital processing technology, the construction of digital receivers, digital terminals, commercial equipment, electrical equipment and optical observation equipment on the site makes the electromagnetic environment of the site extremely complex. [0003] The strength and spectral density of radio frequency interference (RFI) will make the observation results deeply affected by the radio frequency interference and lose the use value. Observations made with single-antenna radio telescopes (continuum or s...

AI Technical Summary

Problems solved by technology

[0005] 1. With the influence of polarization characteristic interference signals, most of the receiver polarizers equipped with radio telescopes are linearly polarized dual-channel output. Through the statistics of the input power of the astronomical observation terminal, the difference in the receiver dual-channel power is large, and the vertical The power of the polarized channel is strong, and with the increase of the pitch angle of the radio telescope, the power of the vertically polarized channel tends to weaken. Therefore, it is believed that the vertically polarized signal from the ground enters the receiving system and affects astronomical observations;

[0006] 2. Due to the impact of sudden interference signals, the interference signals entering the receiving system at a certain time increase significantly, which seriously deteriorates the observation data. The existence of a large number of interference signals makes the observation data in this period of time no longer valid;

Traditional signal-to-noise separation is achieved by comparing the amplitude of each point in the spectrum with a threshold. Frequency points with amplitudes greater than the threshold are considered signals, and frequency points smaller than the threshold are considered background noise. The determination of the threshold is the most difficult, and an inappropriate threshold means Signal extraction accuracy is not high

The method recommended by the International Telecommunications Union (ITU-R) has a narrow frequency band for obtaining spectral noise representation, and its algorithm is not applicable to wideband spectrum in principle

However, the currently commonly used method of manually judging and reading spectrum noise has large errors, and because the radio spectrum noise in the frequency band has fluctuations, it must be manually judged by segments, which greatly increases the workload.

In addition, the acquisition method of spectral noise given in the article "Background Noise Extraction Using Smooth Filtering Threshold Limiting Method [J]. China Radio, 2008(6):75-77." published by Wu Ruijuan and Gong Xiaofeng is only theoretical A solution is given, but in the actual implementation, it is difficult to select the discriminant parameters and determine the size of the frequency band division, which greatly affects the practicability

To sum up, for wideband (narrowband is also applicable) and undulating radio wave environment spectrum, the existing noise and signal separation methods have poor general effect, signal extraction accuracy is not high, and bandwidth applicability is poor

[0013] For this reason, a method for extracting interference signals of broadband spectrum is proposed in the prior art (patent number: ZL201510542046.0.), this patent has the following disadvantages: (1) The adjacent value comparison discriminant value is based on engineering experience and lacks theoretical support. The uncertainty of adaptability and accuracy is large; (2) The adjacent value comparison algorithm is relatively complex, and the determination of the dynamic threshold increases the filtering link, and the calculation speed is slow

Images