Performance Enhancement of the Location and Recognition of a Phi-OTDR System Using CEEMDAN-KL and AMNBP

Yanzhu Hu,Zhen Meng,Xinbo Ai,Yu Hu,Yixin Zhang,Yanchao Shao

DOI: https://doi.org/10.3390/app10093047

2020-01-01

Abstract:Featured Application This research can be applied toward identification and location based on the Phi-OTDR system in fields such as perimeter intrusion, which considers both the calculation efficiency and accuracy in practical environments. Abstract It is commonly known that for characteristics, such as long-distance, high-sensitivity, and full-scale monitoring, phase-sensitive optical time-domain reflectometry (Phi-OTDR) has developed rapidly in many fields, especially with the arrival of 5G. Nevertheless, there are still some problems obstructing the application for practical environments. First, the fading effect leads to some results falling into the dead zone, which cannot be demodulated effectively. Second, because of the high sensitivity, the Phi-OTDR system is easy to be interfered with by strong noise in practical environments. Third, the large volume of data caused by the fast responses require a lot of calculations. All the above problems hinder the performance of Phi-OTDR in practical applications. This paper proposes an integration method based on a complete ensemble empirical mode decomposition with adaptive noise and Kullback-Leibler divergence (CEEMDAN-KL) and an adaptive moving neighbor binary pattern (AMNBP) to enhance the performance of Phi-OTDR. CEEMDAN-KL improved the signal characteristics in low signal-to-noise ratio (SNR) conditions. AMNBP optimized the location and recognition via a high calculation efficiency. Experimental results show that the average recognition rate of four kinds of events reached 94.03% and the calculation efficiency increased by 20.0%, which show the excellent performance of Phi-OTDR regarding location and recognition in practical environments.