Implementation of feature extraction from Rydberg atom EIT-stark spectra of optimized multi-scale wavelet transform and high-precision quantum electric field measurement
Electric field measurement based on Rydberg atom is an emerging development direction, which has attracted attention because of its excellent characteristics such as high precision, high stability, and self-calibration, and also has a large application space in power system. When a non-resonant frequency electric field acts on a Rydberg atom, the frequency shift of the transmission spectrum peak and the electric field strength satisfy a specific mathematical relationship. However, the electric field causes the Rydberg atomic degenerate energy level to split to form an EIT-Stark spectrum containing multiple fine energy level information, which has many difficulties in accurate and online feature extraction of this spectral peak, and affects the accurate inversion of the electric field. In order to solve the above problems, this paper proposes a feature extraction method based on multi-scale wavelet transform to solve the spectral peak feature recognition in the case of transmission peak overlap and transmission peak serious depression. In addition, the scope of feature extraction is limited based on the principle of maximizing variance between classes, and the redundancy of the algorithm is reduced to meet the needs of engineering applications. Compared with the Gaussian fitting method and the multi-scale wavelet transform peak finding method based on mass spectrometry, the feature extraction method proposed in this paper has the advantage in two indicators: the peak false detection rate and the true positive rate. In addition, the method proposed in this paper is integrated into the quantum measurement of electric field, and the error of the measured results is less than 1% after simple parameter compensation for the non-parallel metal plates and the interference of the environmental electric field, which is superior to the existing traditional electric field measurement methods and the conventional EIT-Stark spectral processing method, and also effectively controls the measurement error of the small electric field without significant energy level splitting.