Weak signal data enhancement method

Abstract

The weak signal data enhancement method disclosed by the invention is high in data acquisition precision and strong in anti-interference capability. The method is realized through the following technical scheme: in a data enhancement model, original signal data specifying specific characteristics through a variational self-encoding module to generate a preliminary enhancement signal; mapping the probability distribution density and codes of the original signal data to a hidden variable sample space, and calculating a mean value and a variance of hidden variable samples of the original signal data; the generative network performing random sampling from known obeying likelihood distribution data to obtain hidden variables, and calculating a preliminary enhancement signal obeying Gaussian distribution; the generative adversarial module adopting a generator to carry out increment generation on the preliminary enhancement signal to obtain an optimal generator and a discriminator, the generator generating final enhancement signal data according to the preliminary enhancement signal output by the variational self-encoding module, and then the trained network being utilized to generate an enhanced high signal-to-noise ratio signal, and the probability of outputting a real signal being calculated. and the training module being used for iterative training of the generator and outputting enhanced signal data.

Description

technical field [0001] The invention relates to the field of signal data processing, in particular to a weak signal data enhancement method. Background technique [0002] In recent years, more and more GPS products have been applied to people's lives. However, in harsh environments such as high-rise buildings, indoors, and forests, the GPS signal will be seriously attenuated due to occlusion or reflection, and its carrier-to-noise ratio will usually be low. It is less than 36dB-Hz, so it is difficult for general GPS receivers to capture or track this weak signal. Weak signal detection is a subject that continues to develop with engineering applications. It uses electronics, information theory and physical methods to analyze the causes and laws of noise generation, study the characteristics and correlation of the measured signal, and adopt a series of signal processing methods. method to detect weak signals that are overwhelmed by a noisy background. Weak signal processing ...

AI Technical Summary

Problems solved by technology

There is a strong correlation between the values ​​of deterministic signals at different times; for interference noise, because of its strong randomness, the correlation of values ​​at different times is generally poor

Images