low-energy mobile packet radio networks : routing, scheduling, and architecture

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2000.Includes bibliographical references (leaves 171-176).Packet Radio Networks (PRNETs), which are also called ad-hoc networks, have the capability of fast (and ad-hoc) deployment and set-up, and therefore potentially have several useful civilian and military applications. Building low-energy PRNETs is an important design goal, because the communication devices are typically powered by batteries, and therefore are useless when the batteries are depleted. We choose to look at low-energy PRNETs by focusing on the problem of minimum-energy communication over a PRNET, resolving any related issues or design decisions in a manner consistent with the overall goal of low-energy PRNETs. We conclude that the problem of minimum-energy communication over a PRNET is really a joint routing-scheduling-topological problem. We find the joint problem to be intractable, and therefore propose to solve it by decomposing it, solving each component separately. The resulting solution is not optimal but the degree of suboptimality depends on how the problem is decomposed. Therefore we compare different decomposition methods, and select the one that is likely to yield the best solution to the joint problem. After deciding how to decompose the joint problem, we study the separate components. For the topological problem we decide that nodes should communicate with a limited number of other nodes, referred to as neighbors. We also propose and analyze the performance of a procedure for managing the set of neighbors. For the scheduling problem, we propose a novel and practical class of scheduling algorithms. The routing problem is more complex than wireline routing because of interference and fading. When they are incorporated, routing becomes a non-convex problem; and we overcome this by a novel approach that is non-optimal, but is more robust than the optimal approach.by Hisham Ibrahim Kassab.Ph.D