Abstract
Personal Communication Network (PCN) [1]-[3] is an emerging wireless network that will have a dramatic effect on our industry and life in the future. In a PCN, the covered geographical area is partitioned into a set of cells. Each cell has a base station (BS) to exchange radio signals with wireless mobile terminals (MTs). Whenever a MT moves from one cell to another during a call, this transaction should be handed off to the new cell. The primary purpose of this kind of network is to provide multimedia Personal Communication Service (PCS) for universal access between people or between people and place. General speaking, there are two ways to construct such network. One is to construct it from upgrading the traditional cellular telephone network such as GSM, AMPS, WCDMA, etc. The other is to construct it by designing a whole new architecture such as ATM-based PCN and WDM-based PCN. In this dissertation, we focus our research scope on the latter rather than the former to investigate the related design issue of constructing a Personal Communication Network, which is based on ATM infrastructure or WDM infrastructure. Firstly, in the ATM-based PCN, the design problem of allocating the backbone links among the ATM switches is investigated. Since in an ATM-based PCN, the communication path between a pair of mobile terminals might be elongated due to the mobility of terminals, so we need a better design of backbone links such that the network performance affected by the terminal mobility is reduced as less as possible. We show that this problem is NP-hard. A simple and efficient heuristic is also proposed for this problem. Simulation results show that based on the backbone obtained by the heuristic, the average length of each communication path is kept as short as possible even under the mobile environment. As a result, the number of successfully established paths in the network is increased significantly. Secondly, a whole new architecture of Personal Communication Network, which is based on WDM network, is proposed. Although, recently, ATM-based PCN architecture has been proposed and discussed widely [11]-[15] . However, there are still many shortcomings on this ATM-based PCN architecture. For example, in an ATM-based PCN, connection rerouting must be invoked frequently when hand off happens; it in general takes a longer time. However, by employing the broadcast nature of the WDM star couplers, the proposed WDM-based PCN architecture furnishes a very simple and fast hand off procedure. Thirdly, in the proposed WDM-based PCN architecture, when the mobile terminal moves into the neighboring cell, the so-called wavelength-offset conflict problem might occur. The wavelength-offset conflict problem is produced by the assumption that each base station is equipped with only one pair of tunable transmitter/receiver, this is based on cost consideration. However, if the number of transmitters/receivers equipped with each base station equals to the number of available wavelengths in a star coupler, then the offset conflict will not happen any more. To overcome the wavelength offset conflict problem that might happen in this architecture when mobile terminal roams to a neighboring cell, a wavelength assignment scheme is also proposed. This scheme assigns the wavelength channels to the established connections, including the roaming ones, efficiently so that the wavelength offset conflict rate is kept as low as possible. Simulation results show that compared to the random assignment strategy, the proposed scheme furnishes a much better offset conflict rate.