Application of the joint algorithm of smooth pseudo Wigner-Ville distribution and four-parameter particle swarm optimization to BOTDR

Abstract

Short-time Fourier transform is utilized for the data processing of traditional point-frequency Brillouin optical time-domain reflectometry (BOTDR). This method can improve the measurement speed, but theoretically there is a contradiction between spatial resolution and frequency resolution. To avoid this contradiction, a joint algorithm of smooth pseudo Wigner-Ville distribution (SPWVD) and four-parameter particle swarm optimization (four-parameter PSO) algorithm is proposed to improve spatial resolution, frequency resolution and frequency shift extraction accuracy of Brillouin scattering spectrum. SPWVD can smooth the signal to enhance the frequency resolution by adding a sliding window on time and frequency domain simultaneously when making the quadratic time-frequency transforming. To further improve the fit and the extraction accuracy of the frequency shift, the Brillouin scattering spectra (BSS) processed by SPWVD are fitted to find the Brillouin frequency shift (BFS) using a four-parameter PSO algorithm that takes the gain, BFS, full width at half maximum of the BSS, and linear weight coefficients of Lorentz and Gaussian lines as unknown parameters. According to the experimental results, the BOTDR using SPWVD and four-parameter PSO joint algorithm will cost less operation time since it can get better results with fewer accumulations and it does not need to find proper initial values before curve-fitting operation. Meanwhile, the joint algorithm can improve the spatial resolution and frequency resolution over four and two times simultaneously for BOTDR with 100 ns pump pulse width.