Walking intrusion signal recognition method for fiber fence system

A recognition method based on the gait characteristic for walking intrusion signal is presented. The gait characteristic of a normal walker in the nature state is an average gait period of 1.2s, in which a step period is about 0.6s and a foot touchdown time is about 0.2s. When a person walks fast or runs, the step period is reduced to about 0.4s and the foot touchdown time still keeps about 0.2s. It is included in the vibration signal caused by a walking intruder inevitably. So the detection system output signal caused by a human intrusion is intermittent and periodical. If a sensing system output waveform has a period of 0.3–0.75s and a duration time of 0.15–0.25s, the disturbance source can be adjudged as a human intrusion, not as an animal or other random one. The effectiveness of the proposed method is verified by the experimental results with an in-line Sagnac interferometer fiber fence system and a φ-OTDR intrusion detection system, respectively.