Prediction-Based Call Admission Control Schemes for Wireless Cellular Networks

• Main Authors: Chen-Feng Wu, 吳振鋒
• Other Authors: Liang-Teh Lee
• Format: Others
• Language: en_US
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/39590136584552354323

博士 === 大同大學 === 資訊工程學系(所) === 95 === Due to the rapid progression for wireless technology, the cellular networks are also evolved from the first generation to the third generation and even beyond the third generation further. The subscriber services are gradually changed from voice to multimedia since higher speed of transmission will be provided by newer cellular technologies. Consequently, the multimedia services are getting to become the major trend in next-generation cellular networks, and the Call admission control (CAC) will play the key role for guaranteeing the quality of service (QoS) in wireless cellular networks, owing to the scarce wireless spectrum. Because the user mobility is so indeterminate, the goal which keeps both call dropping probability (CDP) and call blocking probability (CBP) below a lower level simultaneously is more difficult and seldom literatures were focused on such issue. Thus, the appropriate prediction information will be useful for a CAC scheme to meet QoS issues of both CDP and CBP at the same time. There have been so many studies focusing on the prediction of user mobility, and they are fallen into two considerations: the cell approach consists in recording handoff statistics on each cell each time; the user approach consists in computing a roaming profile for each user and recording each visited cell during a call. The design issues of a CAC should be also considered in different scenarios by the same token. Moreover, it will be an important issue for the operation of cellular networks to integrate CAC scheme with the prediction of user mobility produced according to the demand of real situation. In terms of prediction of user mobility, two CAC schemes that are coupled with different prediction methods are proposed in this dissertation. Upon actually using scenarios that may be geography constraint, possession constraint of individual information, and so on, two prediction methods for user mobility are proposed simultaneously. For user approach, the Hidden Markov Models (HMM) concept which is suitable for solving a dynamic situation is applied to predict user mobility then provides information for the correspondingly proposed CAC scheme, and the time series prediction which can predict most information, such as system utilization and CDP, is combined with the correspondingly proposed CAC scheme for cell approach. The SMPL (SiMulation Programming Language) has been adopted for the simulation to evaluate the performance of the proposed schemes. The simulation results of both user and cell approaches show that the proposed CAC schemes are superior to the traditional guard channel schemes in terms of CDP and CBP, and better utilization can be also achieved by the proposed schemes. Indeed, according to the simulation results, the proposed CAC schemes that are combined with the appropriate prediction method of user mobility can do some contributions on wireless cellular networks.