LTE (Long Term Evolution)是3GPP推出的未來4G網路主流的技術，在LTE的網路資源存取技術，3GPP目前對LTE的Downlink採用OFDMA (Orthogonal FDMA)技術，而Uplink則是採用SC-FDMA(Single-Carrier FDMA)的技術。SC-FDMA技術與OFDMA不同之處為SC-FDMA有較低的PAPR (Peak-to-Average Ratio)值，以及分配頻譜資源的方式；SC-FDMA在分配頻譜資源時，可分成連續(Localized)的L-FDMA與不連續(Interleaved)的I-FDMA兩種方式，兩者間的Throughput為L-FDMA較佳，I-FDMA卻有較低的PAPR值[1]，為考量Power Efficiency及行動手持裝置的一些資源限制，通常我們會選擇連續分配頻譜(L-FDMA)的方式。在分配頻譜資源時，為了能讓系統效率提升，我們會讓不同的UE依據其Channel condition和時間的不同，使用不同的Carrier傳送資料，這樣的方法通稱為CDS(Channel Dependent Scheduling)。 現在目前探討這個領域的論文，在研究如何分配資源時，絕大多數都沒有考量到不同網路服務的QoS等級或是其延遲的要求，或是過於簡化。實際上LTE的標準中訂出了數種QoS等級，分為保障Bit Rate下限和不給予保障的兩大類，兩類中又細分成許多不同種類的服務[2][16]，甚至使用者也可以依付費等級再作區分。在考量到不同服務的QoS，分配無線頻譜資源便需要依據Bit Rate和服務的優先順序等分配不同數量的資源給UE(User Equipment)。本碩士論文主要是探討在SC-FDMA技術下，設計出一個更貼近現實環境的演算法，能夠依照不同的使用者等級，來區分所分配的資源，且使網路資源使用效率最大化。¬模擬結果顯示，論文中提出的方法確實能夠達到不同QoS對於頻寬的要求並有效的分配資源。 LTE (Long Term Evolution) is a candidate of 4G network released by 3GPP.In downlink ,LTE uses Orthogonal FDMA (OFDMA) as its multiple access and Single-Carrier FDMA (SC-FDMA) as its uplink multiple access technique. The main difference between OFDMA and SC-FDMA is that the OFDMA has higher PAPR.PAPR of OFDMA values than that of the SC-FDMA. Another difference is the mechanism of resource allocation. In SC-FDMA, it uses two manners to assign resource: localized FDMA (L-FDMA) and interleaved FDMA (I-FDMA).LFDMA performs better throughput and IFDMA has lower PAPR, inversely [1]. In order to provide better resource utilization and consider the power efficiency of the mobile devices, this thesis studies the resource allocation scheme for the uplink with L-FDMA scheme. The proposed scheme considers the channel dependent scheduling (CDS) to allocate resource according to UE’s channel condition and time. Current researches of this domain do not take QoS requirements into account or simplify QoS requirements. In LTE’s Spec [2][16] , there are several QoS levels that correspond to different traffic types. Generally, they can be classified into two types: GBR and non-GBR.And these two types are the traffic models we considered in our study. As the QoS requirements are considered, the proposed scheme shall properly arrange the radio resource to meet the desired bandwidth of each UE according to different priorities and changing channel conditions while maximize the resource utilization. And simulation results show that the proposed algorithm can satisfy different QoS requirements in bandwidth and allocate resource to UE efficiently.
