多重速率載波偵測無線環境之隱含式回饋傳輸速度感知功率控制演算法

Abstract

在 IEEE 802.11 裡面，CSMA/CA 機制用來避免碰撞發生。在傳送封包前，它會先去偵測傳輸媒介是否有其他裝置在使用，再來決定要不要傳輸。為了進一步避免碰撞，還可以使用 RTS/CTS 的機制。但是在無線網路的環境下，一個訊號成功接收與否和所需要的訊號強度與接收端所感受到的干擾強度(即SINR)有關。因此，以 physical carrier sense 方式，將傳送端視為中心點來限制周圍的傳輸，會帶來一些隱藏節點和暴露節點的問題。隨著傳輸速率、傳輸對之間的距離、干擾...等變化，也會導致接收狀況有所不同。因此，限制傳輸的範圍應該隨著環境情況有所改變才對。而像RTS/CTS機制，它同樣使用 physical carrier sense，再加上 RTS/CTS 機制裡的 virtual carrier sense 的特性，其實有可能造成更嚴重的暴露節點問題，所以也應當考量接收端的環境狀態來限制干擾。本論文提出一個RPC(Rate-aware Power Control) 演算法，藉由 busy tone 的機制來改善上述的問題。此演算法是以接收端為中心來做傳輸限制，並且會根據各種傳送速率所需的SINR門檻值以及接收端所受到的SINR值來調整傳輸限制，而傳送端就根據 busy tone 的方法來得知可以使用多少功率來傳輸，最後再利用適應性的功率和速率控制機制來決定適合的功率和速率。此RPC演算法的目的是希望提高網路整體的吞吐量，因此我們使用網路模擬器來測量RPC的性能。藉由模擬結果，可以發現RPC確實能夠提高網路整體的吞吐量。

In IEEE 802.11, when a device attempts to transmit information on the wireless channel, it first detects the channel condition based on CSMA/CA in order to avoid collisions. RTS/CTS mechanisms have been introduced to solve the collision issues. However, whether the receiver successfully receives a packet depends on SINR requirement at the receiver. By using physical carrier sense, some transmitting attempts around the receiver can be reduced, but corresponding hidden and exposed terminal problems occur. In order to reduce collisions at the receiver, we have to take transmitting rate, the distance between transmission pair and interference into consideration. In RTS/CTS mechanisms, by combining physical carrier sense with virtual carrier sense, it is possible that more severe exposed terminal problem will take place. Therefore, in order to reduce the interference level and collision at receiver, we have to consider the environment at the receiver. In this thesis, we propose a Rate-aware Power Control (RPC) algorithm with busy tone support to reduce interference and collisions in carrier-sensing wireless networks. By utilizing busy tone, the receiver informs the neighboring nodes of its SINR requirement, and then the transmitting nodes determine a suitable transmitting power and rate. In this way, other receiving nodes will be able to tolerate this interference within the SINR requirement. The goal of RPC is to maximize the network throughput and reduce the interference at receiver. We use network simulator ns-2 to evaluate the performance of RPC and compare with related mechanisms. Simulation results demonstrate that our RPC protocol improves the system throughput significantly.