Fiber Optic Vibration Signal Recognition Based on GOA-VMD and Weighted Reconstruction

In distributed fiber optic vibration sensing systems, the fiber optic vibration signals are often affected by complex real-world environments, resulting in unstable signal recognition outcomes. To address the question, this paper proposes a recognition method combining the Goose Optimization Algorithm (GOA), Variational Mode Decomposition (VMD), and Support Vector Machines (SVM). First, GOA is used to optimize VMD, enabling the adaptive selection of VMD parameters to obtain more stable Intrinsic Mode Functions (IMFs). Next, correlation coefficients are used to screen the IMFs, retaining the highly correlated IMFs, which are then weighted and combined to reconstruct the signal. Time-domain and frequency-domain features are then extracted from the reconstructed vibration signal. Finally, SVM is employed to recognize signals, and GOA is also used to optimize the parameters in SVM to enhance its performance. Experimental results show that the method based on GOA-VMD and weighted reconstruction achieves higher accuracy, verifying the superiority compared with other recognition methods.