Search

Abstract-This paper proposes a dynamic resource-allocation (DRA) algorithm for packet data services in wireless communication systems based on Hopfield neural networks (HNNs). The resource-allocation algorithm assumes a delay-centric approach in that it maximizes resource utilization of the overall system while minimizing the packet delay. The real-time (RT) working capability of HNN hardware implementation means that a very powerful scheduling DRA algorithm can be obtained. A generic formulation of the algorithm is presented to establish the optimal bit rate allocation. In addition, some illustrative examples of this formulation are given, considering specific wireless communication systems, such as General Packet Radio Service (GPRS) or Universal Mobile Telecommunications System (UMTS). lo be more precise, the performance of the proposed DRA algorithm is evaluated in a realistic UMTS scenario, considering both RT and nonreal-time (NRT) services. To obtain the best resource distribution and fulfill the different quality-of-service (QoS) levels required by RT and NRT services, the new HNN-based delay-centric DRA algorithm is performed twice. Initially, only the RT services are considered, and following this, all the NRT services are taken into account. The results reveal that the proposed DRA algorithm outperforms other reference algorithms in terms of not only average packet delay but the allocated total bit rate as well. [ABSTRACT FROM AUTHOR]

With a small amount of software modification, wireless extension points (EPs), which are now commercially available, can be used to improve the throughput capacity of a wireless local area network (WLAN). An EP is an immobile device that has access to the power supply or is equipped with a high-capacity battery but does not have direct access to the Internet. They wirelessly relay data between the access point and the mobile hosts. In this paper, we investigate the optimal placement of the EPs such that the throughput capacity of an urban rectilineal WLAN can be maximized. Two channel models that are based on the Shannon capacity bound and IEEE 802.11 specifications are studied. We first formulate a solution for the optimal EP placement problem in a general network scenario as a nonlinear programming problem; then, we propose an efficient algorithm that determines the optimal locations of a fixed number of EPs in a rectilineal network. Our results show that, for a wide range of system parameters, the optimally placed EPs can significantly increase the network throughput capacity. Moreover, we study how the number of EPs, transmission power, path loss exponent, channel models, and traffic characteristics affect the optimal EP placement and expected throughput capacity of the network. [ABSTRACT FROM AUTHOR]

In this paper, we present an opportunistic packet scheduling algorithm based on buffer management (OSBM) over the downlink of a packet cellular network. OSBM is a channel-dependent scheduling algorithm with a provable delay bound. It is able to provide differentiated services to both real-time and nonreal-time applications. Particularly, these features of OSBM are achieved through a novel buffer management scheme. Since this buffer management scheme does not involve any complex online computation, OSBM is very efficient and easy to implement in an operational environment. We also present a new analytical approach for performance analysis of opportunistic scheduling in wireless networks based on the proposed concept of effective downlink capacity (EDC). This approach attempts to adapt the service curve tool for deterministic quality-of-service analysis to the wireless environment, and the concept of the EDC serves to bridge the deterministic method and the stochastic nature of the wireless link. Using this approach, the explicit expression of the delay bound of the OSBM algorithm is obtained. Simulation results are presented to demonstrate the effectiveness of the proposed opportunistic scheduling algorithm. [ABSTRACT FROM AUTHOR]

With various wireless technologies developed over the past few years, a ubiquitous and integrated architecture is envisioned for future wireless communication. An important optimization issue in such an integrated system is how to minimize the overall communication cost by intelligently utilizing the avail- able heterogeneous wireless technologies while, at the same time, meeting the quality-of-service requirements of mobile users. In this paper, we first identify the cost-minimization (CM) problem to be NP-hard. We then present an efficient minimum-cost data- delivery algorithm based on linear programming (LP), with various constraints, such as channel bandwidth, link costs, delay budgets, and user mobility, taken into consideration. In case of insufficient bandwidth for communication with the core network, prefetch is employed to fully utilize the wireless-network capacity. If multiple routes are available, a probability-based approach is taken for CM. Extensive simulations are carried out to evaluate the proposed CM scheme. Our results show that the proposed LP approach can effectively reduce the overall communication cost, with small overhead (< 3%) for signaling, computing, and handoff. We expect that minimum-cost data delivery will become imperative for the future heterogeneous wireless networks and the emerging 4G wireless systems. [ABSTRACT FROM AUTHOR]

The focus of this paper is to analyze the throughput performance of TCP Thhoe and Reno on a differentiated services network with correlated wireless losses. An analytical model of TCP on a wireless link with differentiated services (DiffServ) based on the packet train model is developed. The network model is based on the state dependent arrival process. A major contribution of this paper is an analytical model combining the performance of TCP throughput and DiffServ leading to a model that represents an almost complete network scenario. It is shown that DiflServ in a wireless channel with correlated error losses is possible if the prob' ability of the channel changing from a good state to a bad state is less than a particular value. It is also shown that TCP Reno has a higher sensitivity to DiffServ than Tahoe. Our analytical model is validated through a number of simulations. [ABSTRACT FROM AUTHOR]

This paper provides mobility estimation and pree-diction for a variant of the GSM network that resembles an ad hoc wireless mobile network in which base stations and users are both mobile. We propose using a Robust Extended Kalman Filter (REKF) to derive an estimate of the mobile user's next mobile base station from the user's location, heading, and altitude, to improve connection reliability and bandwidth efficiency of the underlying system. Our analysis demonstrates that our algorithm can successfully track the mobile users with less system complexity, as it requires measurements from only one or two closest mobile base stations. Further, the technique is robust against system uncertainties caused by the inherent deterministic nature of the mobility model. Through simulation, we show the accuracy of our prediction algorithm and the simplicity of its implementation. [ABSTRACT FROM AUTHOR]

In this paper, a hybrid resource management system for the uplink in Universal Mobile Telecommunications System wideband code-division multiple access (WCDMA) with components in the Node-B and user equipment (UE) has been proposed. A rate scheduler in the client focuses on average packet delays as a means of abstracting application-specific requirements from the rest of the resource management scheme. It controls uplink transmission through variable spreading gain to optimize resource usage while meeting target delays. Service change requests from the distributed rate schedulers are collectively processed through interservice and intraservice priority queuing in a manner that is shown to exhibit fairness in allocation of resources when cumulative load exceeds system capacity. The performance of the proposed algorithm is explored through discrete-event simulations for three classes of traffic, namely voice, video, and data, over the WCDMA uplink in the presence of short-term Rayleigh fading, automatic repeat request, forward error correction, target transmission delays to meet the respective quality of service, and frame error rate targets in a ‘multicell’ environment. The authors analyze two alternatives for distributed resource management with the UE or Node-B in control of rate scheduling and observe the fairness in resource allocation of both systems. Priority of speech, video, and data traffic is respected and reflected in 95th percentile transmission delays for heavily loaded systems. [ABSTRACT FROM AUTHOR]

Positioning algorithms in cellular networks has become increasingly important as a means of supporting emerging services that require a sufficiently precise estimation of the position of the mobile terminal (MT) associated with a given base station (BS). Currently, even the most sophisticated positioning algorithms require at least three BSs to achieve satisfactory precision. This paper presents a novel algorithm that makes use of a single-BS antenna array to locate MTs in cellular networks. A triangulation technique is utilized and supported by some minimal information about the environment in the BS neighborhood. This algorithm is shown to perform well when operating in a microcellular environment with perfect channel-parameter estimation. The effect of finite resolution of the input parameters is also investigated. The performance is analyzed for a universal mobile telecommunications system microcellular scenario through a three-dimensional deterministic channel model. Finally, the performance of the proposed positioning technique is compared to the well-known location method based on the time-of-arrival measurements at three different BSs. [ABSTRACT FROM AUTHOR]

We study the problem of accurate localization of static or mobile nodes in a wireless ad hoc network using the distance estimates of a group of untrusted anchors within the communication range of the nodes. Some of the anchors may be malicious and may independently lie about the distance estimate. The malicious anchors may also collude to lie about the distance estimates. In both cases, accurate node localization may be seriously undermined. We propose a scheme that performs accurate localization of the nodes in the network despite the presence of such malicious anchors. We also show how to identify most of these malicious anchors. In the case where measurements are error free, we derive a critical threshold B for the number of malicious anchors that can be tolerated in the localization process without undermining accuracy. We also show how to correctly localize a node and identify all the malicious anchors in this setting. In the presence of measurement errors, we propose a convex optimization-based localization scheme that can accurately localize a node, as long as the number of malicious anchors in its communication range is no more than B. Simulation results show that our schemes are very effective. When the measurements are error prone and the number of malicious anchors is no more than B, our scheme localizes a node with an error less than 8% and is also able to identify a significant number of malicious anchors. Our schemes guarantee that a true anchor is not identified as malicious. [ABSTRACT FROM AUTHOR]

Several channel-allocation algorithms have been pro- posed in the literature for traditional cellular networks with static base stations (BS5), while not much attention has been given to channel-allocation problems for cellular networks where BSs are mobile. In cellular networks with mobile BSs (MBSs), the network topology is dynamic due to the mobility of the BSs. This makes the channel-allocation problem in such networks more challenging. Moreover, the MBSs may fail and degrade the performance of the cellular network. Therefore, it is desirable that the channel- allocation algorithm be fault tolerant. Thus, even in the presence of failure of the MBSs, the channel-allocation algorithm should work reasonably well. It is also desirable to design a channel-allocation approach that guarantees that the request from a mobile host (MH) is responded to in a timely manner. In this paper, a distributed dynamic channel-allocation algorithm that addresses these problems in cellular networks with MBSs is proposed. The proposed algorithm is fault tolerant and guarantees that the time to respond to an MH's request for supporting a call is upperbounded. The performance evaluation of the proposed algorithm shows that the proposed algorithm performs well, even in the presence of failure of the MBSs. [ABSTRACT FROM AUTHOR]