Intelligent Mobile Data Mules for Cost-Effcient Sensor Data Collection

Abstract

Sensor networks represent an important component of distributed pervasive infrastructure. A key challenge facing sensor networks is cost-effcient collection of data streaming from these distributed data sites. In this paper, we present a mobile data mule-based sensor data collection approach employing K-Nearest Neighbours queries. We propose a novel 3D-KNN algorithm that dynamically computes nearest sensors spread within a 3D environment around the data mule. The 3D-KNN algorithm incorporates a novel boundary estimation and neighbour selection algorithm to compute the nearest neighbour set. Further, we propose a neighbour prediction algorithm that computes sensor locations within the vicinity of the data mules' trajectory. We simulate the proposed 3D-KNN algorithm using GlomoSim validating its cost-effciency by extensive evaluations. Results of our simulations conclude the paper.