IEEE 802.16e 系統中頻寬請求機制之研究

Abstract

IEEE 802.16e規格中為了支援不同服務之不同服務品質要求，定義了多種頻寬請求機制，透過頻寬請求機制，需要額外頻寬的用戶可以向基地台要求額外的頻寬。其中一種頻寬請求機制是基於競爭的方式，基地台分配額外的資源給用戶端們傳送競爭訊號，用戶們隨意挑選傳輸機會傳送頻寬請求訊號，這些訊號有可能會發生碰撞。為了降低碰撞所發生的overhead，在IEEE 802.16e OFDMA模式中特別設計了展頻碼供頻寬請求機制使用，基地台所分配之頻寬請求資源會先切割成多個傳輸機會,需要請求頻寬的用戶端們會在隨意選擇的傳輸機會中傳送隨意挑選的展頻碼給基地台,一個傳輸機會中允許不同的用戶傳送不同的展頻碼，一旦展頻碼被偵測出，該用戶會被安排到特定的資源位置進行bandwidth request header之訊號交換。本論文針對OFDMA模式下基於展頻碼的頻寬請求機制提出偵測方法，而此偵測方法之計算複雜度會隨著展頻碼的個數增加。為了減少計算複雜度，本論文基於上行通道化提出另一個較為簡單的頻寬請求機制。此機制中一個傳輸機會所需的子載波個數少於展頻碼方法中所需的個數，並且基地台會事先指定傳輸機會給每個用戶。此外，每個細胞有一組獨特碼，屬於同一個細胞的用戶就在基地台所指定的傳輸機會傳送這組碼，基地台只需要偵測傳輸機會是否有用戶使用,而不需要偵測一個傳輸機會中有哪些碼有被用戶使用，此方法在不犧牲偵測率的情況下，大大降低了基地台所需之偵測複雜度。

To support different levels of QoS services, IEEE 802.16e specifies some bandwidth request mechanisms. Using these bandwidth request mechanisms, the Subscriber Stations (SSs) which need some other bandwidth can request bandwidth from the Base Station (BS). One of these mechanisms is contention-based. For contention-based bandwidth request mechanism, the BS allocates resources in the uplink subframe for SSs to send contention signals. The allocation is divided into transmission opportunities (TOs). Each SS randomly select a TO and send contention signals which may result in collision. To reduce overhead, IEEE 802.16e specifies a CDMA-based bandwidth request mechanism for OFDMA-PHY. SSs which want to request bandwidth send CDMA codes as contention signals in their selected TOs to the BS. More than one CDMA code are allowed in a TO. In this thesis, a detection method is proposed for this CDMA-based bandwidth mechanism. To reduce computation complexity, we also propose a simpler bandwidth request mechanism. In the proposed method, the number of subcarriers in a TO is less than that in the CDMA-based method and each SS is pre-assigned a TO by the BS. Besides, each cell has a unique code. SSs from a cell transmit the code in the pre-assigned TO. Thus the BS does not need to detect all the codes for a TO. Without sacrificing detection probability, this method greatly reduces computation complexity.