OFDMA系統在分配頻寬時，除須考慮到通道品質情況(Channel Condition)分配資源以提供有效的服務品質(QoS)需求外。在OFDMA系統下分配頻寬時，也需要考量到訊務塊(Traffic Burst)多寡對訊框中MAP大小所造成之頻寬額外負擔(Overhead)，才能使無線頻寬資源分配更有效率。在文獻[2]中提出，OFDMA系統的MAP在下載鏈路(Down Link, DL)上會耗掉20%-60%的頻寬資源，提醒OFDMA系統會有著過大Overhead的關鍵性問題存在。由於此問題為初步被提出，尚未有系統化方法之討論，因此本論文提出一個兼顧無線通道品質與調變選擇之訊務塊配置演算法，以降低MAP所造成之額外頻寬負擔的影響。 我們所提出的演算法根據用戶端(Subscriber Station)之頻寬需求及其通道適合之調變方式，並在考慮計算複雜度前題下，提出利用調適性邊線來界定可分配目標區(Target Site)的觀念，儘量以矩形區塊方式來分配用戶端所需求之頻寬，以減少訊務塊數量，進而有效的減少MAP欄位所需之位元數。除此之外，演算法對於是否降低調變以有利矩形區塊之對應，也考慮在最佳系統疏通量(Throughput)情形下，提出有效之法則來判斷最底限的調變。模擬的結果顯示，本論文中所提出的演算法能在QoS的考量下，有效改善OFDMA之額外負擔問題。與未考量MAP大小的Mapping演算法比較，模擬結果顯示所提出之方法可提升約16.5%之系統效能。 In the OFDMA system, the channel condition shall be considered as the basic requirement of resource allocation for the provisioning of QoS services. In addition, the size of MAP, resulting from the increase of the number of traffic bursts, shall also be well considered to reduce its overhead so that the valuable bandwidth of wireless can be effectively utilized. It was illustrated in the reference [2] that the MAP overhead of OFDMA system will exhaust 20%-60% bandwidth of down link and it also pointed out that this is a critical problem of bandwidth utilization in OFDMA system. As this issue has not been well discussed before, a noval traffic burst allocation algorithm, which considers the channel quality, coding and modulation, is proposed in this thesis to reduce the influence of the extra bandwidth burden in MAP. The proposed burst mapping algorithm considers not only the bandwidth requirements of Subscriber Stations (SS) and applicable channel modulation, but also the time complexity. And, in order to map the traffic bursts in rectangle blocks more effectively, the concept of “Target Site” with flexible boundary adaptation is applied. Besides, if a traffic burst can not be fit into a rectangle block, an appropriate degradation of the coding and modulation level is derived by considering the optimal throughput. The simulation results show that, under the QoS constraints, our algorithm can effectively improve the extra bandwidth burden caused in OFDMA system. When Comparing to the existing normal mapping algorithm, the simulation result shows that our algorithm can improve about 16.5% of the system performance.
