提出了运用信号自相关矩阵特征值(EVD)/奇异值分解(SVD)的谱估计方法对法布里-珀罗(F-P)传感器的频分复用系统进行实时、高分辨率和低串扰的信号解调。从理论上分析了该方法对法-珀传感器复用信号解调的可行性,并在短采样数据长度条件下对两个传感器的并联复用信号进行了解调以及串扰的实验研究。实验表明,与离散傅里叶变换法(DFT)、Pisarenko等算法相比,该方法分辨率更高,在两传感器的腔长差低至20 μm时,运用该算法仍可以实现准确的解调,而因串扰引起的应变误差小于±12 με;此外,短的采样数据长度决定了该算法的运算量较小,信号处理速度较快。因此,该方案在大容量准分布式传感网络系统中将具有极大的潜在应用价值。

A new frequency estimation method with high-frequency resolution and low-crosstalk properties based on the eigenvalue decomposition (EVD) and singular value decomposition (SVD) of autocorrelation matrix for frequency-division-multiplexed Fabry-Perot(F-P) sensor network is proposed. Embarks from the basic principle of F-P interferometer, its feasibility on demodulating multiplexed signals is demonstrated. With a short sampling length, the shunt multiplexing signal of two sensors is demodulated and the crosstalk is analyzed in experiment. It is found that the proposed method owns a higher resolution compared with diverse Fourier transform (DFT) and Pisarenko methods. The accurate demodulation is carried out with stress error induced by crosstalk less than ±12 με when the cavity length difference is only about 20 μm. In addition, the signal processing speed is significantly increased as the number of data sequence could be reduced to 400, which makes this algorithm stronger for real-time demodulation. Therefore, such method proposed has great application potential in high-multiplexing-capability sensing networks.