Constructing load-balanced Degree-constrained Data Gathering Trees in Wireless Sensor Networks

The tree-based routing topology is widely used in various Wireless Sensor Networks (WSNs) to provide periodic data gathering. Sensors gather data and then report to the sink after data aggregation. Since sensors are usually energy limited in WSNs, it is always important to conserve node energy and extend the network lifetime that defined as the time elapsed since the network starts until the first sensor runs out of energy. Extensive research attempts to solve it as a load balancing problem aiming at balancing energy depletion among sensors. This paper studies load balancing on the tree-based data gathering networks (i.e. data gathering trees). We find that, in a Degree-constrained Data Gathering Tree (DcDGT), when the tree´s height is large enough, the data traffic and load of a node are mainly determined by the distributions of its descendant nodes whose distances to it are within a certain number of hops. Based on this finding, we propose a load estimation model and further give a distributed load-balanced DcDGT construction algorithm. Simulation results demonstrate that our algorithm effectively prolongs the network lifetime and outperforms existing approaches.