Identification of trespasser from the signatures of buried single mode fiber optic sensor cable

It is common awareness that occasionally wild animals forage into human habitations in search of food and humans forage into the forest for livelihood and poaching. This trespassing into each other's territory often leads to human-animal conflict. Further, the human presence and actions in the forest would lead to the destruction of the forest and degradation of the forest habitat for the wild animals. This paper presents the work of the authors in identifying the intruding animals/humans based on the characteristic signature of the animals obtained from the buried fiber sensor. The buried fiber sensor is sensitive for about two meters on either side of the fiber and provides a characteristic signature when an intruder walks over the ground where the fiber is buried. The identification process of the intruder is invariant to the weight of the animal, speed of the animal, type of the soil, the depth at which the fiber is buried, the environmental conditions near the fiber. Since the cable and associated electronics are buried in the earth, this fence is not visible to anyone. Algorithms for identification of the individual species (elephant, tiger, and human) needed training. A number of signatures each of the animal categories were collected from zoos and forests. Using this data the system was trained to recognize elephant, tiger, and human. Using this trained model the animal signatures were classified. Three different algorithms were used to classify the trespassers. The identification of trespasser is based on majority output of all three algorithms. The efficiency of the algorithm is 94.7% for human, 87.4% for tiger and 88.7% for elephant.