無線網狀網路之空間淨空重用與頻譜效益分析

Abstract

使用無線網狀網路（wireless mesh networks）作為新一種無線寬頻存取技術時，使用者預期該網路應有龐大的網路傳輸容量（network capacity throughput）。由於無線傳輸廣播的特性，在此種網路下，同時傳輸的無線訊號之間的相互干擾（interference）會對整體網路傳輸容量造成很大的影響，對一個無線傳輸而言，訊號干擾愈嚴重的話，其可使用之傳輸速度就愈低。 對於無線網狀網路而言，其網路傳輸容量由傳輸速度高低與同時傳送資料的基地台數量這兩個因素來決定。因此，提升網路傳輸容量最好的辦法就是提高傳輸速度以及提高同時傳送資料的基地台數量；然而，無線傳輸在高的速度之下較無法抵抗干擾，因此當提高傳輸速度時，我們無可避免的必須要減少同時傳送資料的基地台數量。因此對於（1）提高傳輸速度以及（2）增加同時傳送資料的基地台數量這兩個選擇我們只能擇其一，無法同時達到兩者。 在本論文中，假設使用切時分工多重存取（time division multiple access）、分散式資料傳輸排程、以及實體干擾模型之無線網狀網路下，我們探討這兩種選擇結果對網路傳輸容量的影響並找出達成最大網路傳輸容量的選擇。

The wireless mesh network is expected to provide large transmission capacity in order to serve as a complementary broadband wireless access technology. The interference plays an important role to the network capacity throughput of wireless mesh networks because when there are multiple concurrent wireless transmissions in a mesh network, the interference caused by the broadcast nature of wireless transmissions decreases the sustainable spectral efficiency (transmission speed) of each individual transmission. For a wireless mesh network, the network capacity throughput is determined by two factors: the individual transmission spectral efficiency and the number of concurrent transmissions. Therefore, the best way to increase the network capacity throughput is to increase both the individual transmission spectral efficiency as well as the number of concurrent transmissions. However, if we increase the individual transmission spectral efficiency, inevitably we have to reduce the number of concurrent transmission because we must keep interference low to make higher spectral efficiency feasible. Thus the choice between (1) increasing the transmission spectral efficiency and (2) increasing the number of concurrent transmissions is a trade-off and is critical to the network capacity throughput. In this paper, we explore the tradeoff between these two choices in a TDMA (time division multiple access) based distributed data access scheduler wireless mesh network with physical interference model and find the optimal design that achieves highest network capacity throughput.