標題: 利用類神經乏晰方法之無線ATM網路之允諾呼叫控制
Neural Fuzzy Approach for Call Admission Control in Wireless ATM Networks Using PRMA Protocol
作者: 陳義昇
Chen, Yih-Shen
張仲儒
Chung-Ju Chang
電信工程研究所
關鍵字: 允諾呼叫控制;類神經乏晰;無線ATM網路;call admission control;neural fuzzy;wireless ATM network
公開日期: 1996
摘要: 隨著通訊科技上的進步,未來寬頻無線網路正在迅速地發展。另一方面 ,非同步傳輸模式網路的發展也愈趨成熟。在資訊發達的今日,非同步傳 輸模式網路可以有效的把多種通訊服務整合在一起。因此許多的研究都集 中研究於無線非同步傳輸模式網路。無線非同步傳輸模式網路是以非同步 傳輸模式網路為骨幹網路,並且在無線介面中提供整合性的寬頻服務。因 為無線頻寬資源的稀有性,以及提供具有服務品質保證的服務,無線非同 步傳輸模式網路需要對系統資源做有效的管理。 在這篇論文中我們將提 出一個無線非同步傳輸模式網路的架構。和有線非同步傳輸模式網路不同 ,無線非同步傳輸模式網路需同時考慮無線介面以及有線介面的服務品質 。首先,我們針對低移動性的無線非同步傳輸模式網路提出一個類神經乏 晰呼叫允諾控制器。接著,我們加入對使用者移動性的考量。因此,藉由 加入新的乏晰預估器以對周圍頻寬進行監測,我們把新的類神經乏晰呼叫 允諾控制器應用於高移動性的無線非同步傳輸模式網路上。從系統模擬的 結果可以看出,我們所設計的類神經乏晰呼叫允諾控制器確實可以有效提 供倔有服務品質保障的服務。同時,這過類神經乏晰呼叫允諾控制器也可 有效率的對系統資源進行管理。 with the innovation of many new communication technologies, wireless broadbandsystem will be available in the future. ATM network, which flexibly and efficientlyntegrated multiple kinds of communication services, are considered as the backboneof the construction of B-ISDN. Recently, Many researches are focus onwireless ATM (WATM). The WATM networks not only provide wireless access to the broadband network but support seamless services in a single network. In this paper, we propose the architecture of WATM. Because of the preciousnessof wireless frequency spectrum and the user mobility, the WATM needs an appropriate resource management scheme to achieve high system utilization.Unlike the wireline ATM networks, the user mobility and the QoSs of air interfaceand the wireline networks are needed to take a serious consideration. We first designa neural fuzzy call admission controller (NFCAC) for the resource managementwithin a cell in the WATM network. The NFCAC is a two- layer neural fuzzycontroller. In the first layer, fuzzy estimators measure and monitor the performanceof different parts in the WATM. A neural fuzzy controller in the second layermake a CAC decision. Since we appropriately choose the input linguistic variables,the controller can have robust and stable operations. Then, we further consider the resource management for the overall WATM network. We design a neural fuzzycall admission controller (WATM-NFCAC) by the same methodology. By adding afuzzy estimator to predict the loading of the neighboring cells, the WATM-NFCACcan easily manage the user mobility. The simulation results show that theWATM-NFCAC can adaptively reserve suitable capacities for the handoff calls withoutdramatically increasing new call blocking probability. Also, the WATM-NFCAC canachieve high system utilization while keep all the QoS contracts that claimed.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT850436047
http://hdl.handle.net/11536/62125
顯示於類別:畢業論文