11-05-2013, 03:51 PM
Proportional fair switched channel scheduling Schemes for multiusers with optimal threshold
Proportional fair switched.docx (Size: 34.94 KB / Downloads: 19)
Abstract
Multiuser switched-diversity scheduling schemeswere recently proposed in order to overcome the heavy feedbackrequirements of conventional opportunistic scheduling schemesby applying a threshold-based, distributed, and ordered schedulingmechanism. The main idea behind these schemes is thatslight reduction in the prospected multiuser diversity gains isan acceptable trade-off for great savings in terms of requiredchannel-state-information feedback messages. In this project work, we characterize the achievable rate region of multiuser switched diversity systems and compare it with the rate region of full feedback multiuser diversity systems. We propose also a novelproportional fair multiuser switched-based scheduling schemeand we demonstrate that it can be optimized using a practicaland distributed method to obtain the feedback thresholds. We finallydemonstrate by numerical examples that switched-diversityscheduling schemes operate within 0.3 bits/sec/Hz from theultimate network capacity of full feedback systems in Rayleighfading conditions.
Introduction
The concept of multiuser diversity (MUD) has been wellstudied in the literature and exploitedin the design of channel-aware “opportunistic” schedulingschemes that control in a dynamic way how the users accessthe shared air-link resources in wireless systems. This conceptwas originally initiated in aresearch work where it was shown that in orderto maximize the sum capacity (bits/sec) of the network; weshould always schedule the user with the best instantaneouschannel quality. The design of opportunistic schedulers hasbeen further studied in the literature taking into considerationkey factors such as fairness among users and maintaining thequality-of-service (QoS) constraints.
In virtually all modern wireless communication systems,explicit training sequences (i.e. pilot signals) are used toenable the receivers to measure and/or predict the instantaneous channel conditions in order to use it in thecoherent detection of the transmitted signals. Opportunisticschedulers that are capable of exploiting the full MUD gainsare based on having continuously-updated channel state information(CSI) of all back-logged mobile users in the networkat the central scheduler (i.e. at the base station). Thus, allmobile terminals inform the central scheduler about their CSIusing explicit feedback messages. As a result, a considerableportion of the air-link resources and a significant share ofthe battery energy of the mobile terminals are used for theCSI feedback instead of useful data traffic. This fact hasmotivated many researchers to examine the feedback load ofopportunistic scheduling schemes1 and to search for alternativeschemes which can trade off some of the MUD gains forconsiderable savings of the feedback load.
Aim and Objective
The main goal of this research work or project work is to develop a noble architecture or design of Multiuser switched diversity scheduling scheme that can accomplish the following objectives:
1. Obtain the fairness in Scheduling scheme
2. Design a system in which a single radio or air link resource can be used for Multi user communication scenario.
3. In spite of conventional selection based scheduling here in this research work, a switching based scheduling scheme has to be obtained that may perform better than the existing systems.
4. A comparison of MUSwiDschemes with full-feedback (MUSelD) opportunisticscheduling schemes is needed to evaluate how much ratewe lose due to the feedback savings.
5. Compare the developed system output with existing full feedback multiuser diversity scheduling system
Existing System
This concept of Multiuser diversity scheduling or “opportunity scheduling” was originally initiated in a proposal where it was shown that in order to maximize the sum capacity (bits/sec) of the network, we should always schedule the user with the best instantaneous channel quality. In many of the research work the design of opportunistic schedulers has been further studied taking into consideration of the dominant key factors like fairness among users and maintaining the quality-of-service (QoS) constraints etc. Considering the quality of service as well as network efficiency, it can be said that a number of parameters are there which are to be optimized so as to achieve a better QoS and system performance. Majority of communication techniques employ individual air link or air resources (channel) to send the feedback signals or CSI. This causes the unnecessary use of air link and thus the system throughput is reduced. But At present there isno general theory of single or multiuser wireless feedbackcommunication networks that can effectively minimize the number of feedback messages and thus preserving a significant amount of air link resource as well as power.
Proposed system
In the proposed system we provide a comprehensive study to answerthe aforementioned technical challenges. Furthermore, we aimin this work to persuade that MUSwiD scheduling systemsare actually attractive options for practical implementation inemerging mobile broadband communication systems. Towardthis end, we take the following steps; we provide detaileddiscussions to enhance our understanding about the attributesof the system and how to optimize its performance. In particular,we characterize the achievable rate region of MUSwiDsystems. Also, we show that the achievable rates in MUSwiDsystems are comparable with selection-based systems althoughthey are significantly more economic in terms of CSI feedbackload. Furthermore, we propose a novel MUSwiD schedulingscheme that achieves the proportional fairness criterion, which is preferable for practical implementation.
We show that this can be achieved by proper per-user thresholdoptimization based on the objective function of maximizingthe sum of the logarithms of the achievable rates. Wedemonstrate that our proposed scheme has a special interestingfeature that the solution of the corresponding optimizationproblem yields independent equations for each user, andhence the threshold optimization can be decentralized, whichovercomes the centralized optimization challenge.