Department of Electrical and Computer Engineering
Ph.D. Thesis Abstract
Capacity of an Infostation System
An Infostation system consists of a network of wireless ports providing discontinuous coverage. In these coverage zones, the conditions for communication are such that transmission at high bit rates is possible. Consequently, we may be able to download large amounts of information even if the mobile terminals spend little time within coverage.
In this dissertation we define several models for an Infostation system and evaluate the performance in terms of capacity, throughput, and delay. The models we consider are the periodic model, the singleInfostation model, and cellular models.
The periodic model studies the case where Infostations are regularly spaced, such that a probability distribution for the mobileInfostation distance can be found. With this distribution it is possible to derive the power control policy to maximize the average capacity in the informationtheoretic sense, or similarly to maximize the BPSK cutoff rate.
The singleInfostation model looks at what is the maximum amount of data that can be transmitted to an isolated Infostation. We find the throughput when continuously adapting the symbol rate, and the optimum rates and coverage radii when restricting to a finite number of rates. We also study the effect of varying the modulation scheme.
In the cellular models, we locate the Infostations in a cellular layout, both in one and two dimensions. We find what is the worstcase SIR as a function of coverage and determine possible operating points according to modulation and frequencyreuse cluster size. Next, we model the Infostation as a queueing system with reneging and study throughput and delay. The results show that for one dimension the best choice is the highest level modulation with the smallest cluster size (which gives the smallest coverage area). For twodimensions there is a tradeoff, such that for small loads a fuller coverage is preferred.
We also consider the introduction of Infostations in an existing cellular system, taking advantage of unused excess capacity. Still using a cellular layout, we analyze the uplink for a static situation, in which the load perceived by the Infostation depends on the area covered. We derive the optimum coverage radius for a given data rate and a maximum throughput constraint. Then we extend the result to the case of two data rates, solving for the corresponding optimum radii that define the coverage areas in which we should use each rate.
Ph.D. Dissertation Director: Professor Roy D. Yates
Download the Thesis: PS