On The Efficient Data Gathering In Large Scale Autonomous Bandwidth-Limited Sensor Networks

In this paper a novel data gathering framework for resource-constrained sensor networks is introduced and evaluated. The introduced framework consists mainly of two phases. Initially, due to the lack of infrastructure, a simple bridging technique is used to create a multi-hop tree rooted at the collection center. In the second phase, a distributed and probabilistic method is applied by the various sensor nodes to perform data aggregation based on their position (level) on the data gathering tree. The adopted aggregation approach aims at utitizing the available limited resources efficiently and effectively reducing significantly the network traffic, and as a result shortening the delays at the intermediate nodes and reducing the corresponding collisions and energy wastage in data transmission. The performance gains that can be achieved by the proposed data aggregation framework are evaluated via modeling and simulation, under different aggregation scenarios and traffic loads.