在LTE直接通訊中資源保留的效能分析

Abstract

最近以鄰近為基礎(proximity-based)的應用服務日趨增加，這些應用服務主要是探測(discovery)鄰近有相同應用服務的裝置，並進一步的交換資訊。雖然目前一些通訊技術可提供這類的應用服務，例如：藍芽(Bluetooth)、WiFi Direct等，但都有傳輸距離短、傳輸速率低、訊號容易受到干擾等缺點。因此，鄰近服務(proximity-based services，ProSe)在3GPP最近幾次會議中成為相當熱門的議題，希望藉由提升現有的3GPP標準技術以全面支援這類型的應用服務，並且可用於公共安全(public safety)的領域當中。雖然直接通訊能夠帶給使用者更大的便利性，也能降低系統不必要的負擔，但因為頻譜是有限的，如果大量的D2D使用者加入到系統中時，很有可能會大量影響原本在系統內的cellular使用者。因此本論文主要目的在於探討不同的D2D使用者數量對於原本在系統內cellular使用者的影響，並且針對這種影響提出一套改良的演算法。模擬結果證明我們提出的演算法能使cellular使用者保有一定程度的服務品質，另外我們也發現遺忘因子的值對於資源分配的公平性有很大的影響。

In recent years, proximity-based applications have become increasingly popular. One of the major purposes is to discover the devices installed the same applications such that information can be exchanged between these devices. Some communication technologies, such as Bluetooth and WiFi Direct, can support such applications. However, there are some weaknesses such as short communication distance, low transmission rate and susceptible to interference. Therefore, the 3GPP is now actively discussing on how to provide proximity-based services (ProSe) by using LTE. How-ever, the throughput of cellular users will be affected when a large number of D2D users join to the networks because of the limited spectrum. Therefore, this thesis aims to explore the impact of different number of D2D users for cellular users. In this thesis, we propose an effective algorithm to solve this problem. The simulation results show that the proposed algorithm can guarantee the cellular user’s throughput to a certain degree. The results also indicate that the forgetting factor has a significant im-pact on the fairness of resource allocation.