Adaptive sampling for node discovery: Wildlife monitoring & sensor network

Searching for the next hop node in mobile sparse wireless sensor networks for data exchange is a challenging task. This involves frequently sending radio beacons that drain battery power and reduces the life of the sensor node. This work proposes a novel energy efficient approach of adaptively sampling the network connectivity. The adaptive sampling starts with random sampling of the network to collect the accelerometer data related to the demographic distribution of the animals. The collected accelerometer data is used to train an Artificial Neural Network (ANN). This then predicts the timing for future sampling. This prediction mechanism reduces the number of beacons transmitted, thereby improving the battery life of the sensor node. The simulation results show that the approach offers one sixth reduction in the required energy for communication. This should significantly improve the operational life of the nodes.