论文部分内容阅读
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20μm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±12με;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 spectral estimation method based on EVD / SVD is proposed for the real-time, high-resolution and low-crosstalk frequency-division multiplexing systems of Fabry-Perot (FP) Signal demodulation. The feasibility of this method for demodulating the signal of the Fabry-Perot sensor is theoretically analyzed. The parallel demodulation of the two sensors and the crosstalk of the demodulator are studied under the condition of short sampling data length. Experiments show that this method has higher resolution than the DFT and Pisarenko algorithms. When the cavity length difference of two sensors is as low as 20μm, the algorithm can still achieve accurate demodulation. However, the strain error caused by crosstalk is less than ± 12με. In addition, the short sampling data length determines that the algorithm is less computationally intensive and the signal processing speed is faster. Therefore, this scheme will have great potential application value in the large-capacity quasi-distributed sensing network system.