Research on identification of multiple target signals based on φ-OTDR

Abstract

Phase sensitive optical time domain reflectometer (φ-OTDR) is widely employed in pipeline safety early warning, perimeter security, Pipeline inspection gauge (PIG) positioning and tracking, etc. Generally, equipment is applied in outdoor environments. There is a high probability that multiple intrusion signals exist simultaneously at a common position. Therefore, it is significant to unmix and identify each signal. In this paper, a multi-target signal recognition method based on φ-OTDR is proposed. The basic principles of φ-OTDR system and fixed point algorithm based on negative entropy maximization (FASTICA) are briefly introduced. The blind source separation and recognition of the mixed signals of hammer and motor are carried out, which proves the effectiveness of the method. In order to further quantitatively evaluate the effect of separation, the feature extraction of hammer signal and motor signal is carried out from two aspects respectively from time and frequency domain. The feature vector describing signal characteristics in terms of both time and frequency domain is proposed and used as the evaluation criterion of the separation performance. The experimental results show that the separated signal basically conforms to the feature vector, and the time domain feature is the most stable, with a maximum error of only 3.08%.