| Extended Abstract
Ultra-wideband (UWB) systems transmit and receive signal through an extremely short-duration burst of radio frequency (RF) energy, typically a few tens of picoseconds to a few nanoseconds in duration. The resultant waveforms have extremely wide frequency band and it is difficult to determine the actual RF center frequency (i.e., “carrier-free”). The amount of spectrum occupied by a UWB signal, i.e. the bandwidth of the UWB signal, should be at least 25% of the center frequency. Since UWB waveforms are of such short time duration, they have some rather unique properties and have found both military and commercial applications, including tactical handheld and network LPI/D radios, non-LOS LPI/D groundwave communications, precision geolocation systems, high-speed wireless LANs, collision avoidance sensors, intelligent tags, among others [2, 3, 5].
In this paper, we consider UWB-based ad hoc networks. Within such a network, there is a group of source-destination communication pairs, which we call sessions. For each session, data is generated at the source node and must be relayed to its destination node, typically via multiple hops. The objective is to maximize the total utility, where a session’s utility is defined as the log function of its data rate [4].
Clearly, this optimization problem involves issues from different layers (i.e., link layer and network layer). Such large optimization space brings in some unique challenges. Due to interference from neighboring links, a change of power level on one link will produce a change in capacity in all neighboring links. As a result, capacity-based routing problem at the network layer is deeply coupled with link layer problems such as scheduling and power control. Inevitably, the problem formulation will involve optimization variables from different layers (or cross-layer).
This problem was first attempted by Radunovic and Le Boudec in [4], where they formulated the problem into a nonlinear programming (NLP). Due to its complexity, they did not solve the optimization problem directly. Instead, they studied simple cases and attempted to extend the results from these simple cases to network with general topology through heuristic approach. More discussions on [4] will be given in the paper. It turns out that claims made on a general network via such heuristic approach are problematic (sometimes incorrect) as we shall discuss
throughout the paper. As a result, a formal mathematical solu |
| Dr. Sastry Kompella is a researcher in the Information Technology Division of U.S. Naval Research Laboratory, Washington, DC. Dr. Kompella obtained his Ph.D. degree from Virginia Tech, Blacksburg, VA in 2006. His research intersts include wireless networks, cross-layer optimization, and video communications. |