The elephant and human conflict is a major problem in Sri Lanka. One way of preventing elephant depredations is by detection of a rogue elephant before it reaches a village or farmed areas through an early warning system so that people can drive the elephant away before it causes damage to their crops or property. Being very intelligent animals, elephants encountering such a new system, often adapt to it very quickly.
Examples observed in Sri Lanka include elephants toppling trees on the electric fences to make them ineffective and digging up and destroying US Army surplus miniature seismic systems installed to detect crop-raiding elephants. Therefore, an adaptive approach is required that is based on multiple methods to detect elephants and that could randomly vary the actuation and this way prevent that the elephants get used to the system.
As an alternative strategy for electric fences and other fix detection systems we propose the use of an Adaptive Sensor Actuator System for Elephant Tracking (ASSET) for elephant detection and early warnings. The system will exploit the recent advances in flying robotic (quadcopter) and wireless sensor network (WSN) technology, which enables monitoring of the environment at low cost. The quadcopter chosen for this project is a Parrot Ar Drone 2.0 from Apple. It will be equipped identify the elephants.