Energy efficient data gathering in multi-hop hierarchical wireless ad hoc networks

This paper studies the problem of data gathering in hierarchical wireless ad hoc networks. In this scenario, a set of wireless devices generates messages which are addressed to the base station. As not all nodes can reach the base station through a direct transmission, messages are relayed by other devices in a multi-hop fashion. We consider data gathering without aggregation, i.e. all the generated messages are required to reach the base station - this is in contrast to the well studied problem of data gathering with aggregation, which appears to be significantly simpler. The above scheme may have poor performance in wireless networks with hierarchical architecture. The devices in the layer closest to the base station form a bottleneck in terms of energy consumption as they experience a very high volume of forward requests. Eventually, these nodes will be the first to run out of their battery charges which will cause connectivity losses. In this paper we focus on prolonging the network lifetime of hierarchical networks through efficient balancing of forward requests, which is NP-hard. We develop an approximation scheme which produces a data gathering tree with network lifetime which is at most k times less than the optimal one, where k is the number of hierarchical layers. Our results are analytically proved and validated through simulations.