Abstract:

Cooperation between the nodes of wireless multihop networks can increase communication reliability, reduce energy consumption, and decrease latency. The possible improvements are even greater when nodes perform mutual-information accumulation, e.g., by using rateless codes. In this paper, we investigate routing problems in such networks. Given a network, a source and a destination, our objective is to minimize end-to-end transmission delay under a sum energy constraint. We provide an algorithm that determines which nodes should participate in forwarding the message and what resources (time, energy, bandwidth) should be allocated to each. Our approach factors into two sub-problems, each of which can be solved efficiently. For any node decoding order we show that solving for the optimum resource allocation can be formulated as a linear problem. We then show that the decoding order can be improved systematically by swapping nodes based on the solution of the linear program. Solving a sequence of linear program leads to a locally optimum solution in a very efficient manner. In comparison to the cooperative routing, it is observed that conventional shortest-path multihop routings incur additional delays and energy expenditures on the order of 70%. Since this initial solution is centralized, requiring full channel stat information, we exploit the insights to design two distributed routing algorithms that require only local channel state information. We provide simulations showing that in the same networks the distributed algorithms find routes that are only about 2-5% less efficient than the centralized solution.