具曼哈頓移動模型性質的無線網路節點其連線性質之研究

Abstract

在無線行動隨意網路中，由於網路連線與拓樸狀態隨時都在改變，進而影響無線網路上繞徑演算法的表現。尤其是不同的無線節點移動模式，往往會對繞徑演算法造成不同的影響。曼哈頓移動模型就是一種常見的節點移動模型。在此模型中，所有的無線節點在格網地圖上移動，進而影響到無線網路上的網路連線與拓樸狀態。在這篇論文裡，我們建立了具曼哈頓移動模型移動性質的無線行動隨意網路，其連線生命週期的數學模型。當兩個無線網路節點互相位於對方的傳送範圍內，此兩無線節點間即建立起一條暫時的連線。此連線會受到節點的移動模型、節點的速度、以及節點的傳輸半徑所影響。因此，具有曼哈頓移動性質的節點，其連線生命週期必然與具有其他移動模型的節點有所差異，而有所必要建立起其獨特的數學模型。

由於在曼哈頓移動模型中，每個無線節點都被限制在一個格網拓墣上。因此，其兩個移動無線節點之間的連線型態可以分成三個獨立的子連線型態：節點相對移動型態、節點平行移動型態，以及節點垂直移動型態。由於這三種形態是獨立的，是故可以利用機率數學工具分別對每個子連線型態分析。爲了驗證數學工具所提供的結果，我們可以藉由NS2網路模擬器來加以印證所提出的具曼哈頓移動性質其無線節點間平均連線生命週期的數學模型。而其結果也與我們所提出的數學模型相當一致。

Since multi-hop mobile ad hoc network (MANET) contains a set of wireless mobile nodes forming a temporary network, the topology of it is strongly affected by the mobility model of nodes. In previous works, there are a great diversity of mobility models have been presented, such as Random Waypoint Mobility Model (RWMM), Manhattan Grid Mobility Model (MGMM), and Freeway Mobility Model (FMM) etc. These works agree by mere coincidence that different mobility models result in different network topology, and then influence the performance of MANET routing protocols.

If the trajectory of a node is confined to a grid topology, we call this node adheres to Manhattan Grid Mobility Model (MGMM). In megalopolis, MGMM is an important mobility model, and lots of objects following MGMM can be enumerated, such as the cars in urban, the people in department stores etc. According to previous works and our knowledge, MGMM has different attributes of MANET from the other mobility models. Moreover, the expected link life time (ELLT) is one of the significant parameters of MANET. When a node wants to transmit information to the other node directly, these two nodes have to exist in each other’s transmission range for a period. This period is called the link life time, and ELLT is the expected value of LLT. The goal of this work is mainly to find out the ELLT of the MANET following MGMM by using mathematical tools. According to our observation, MGMM can be separated into three independent cases: the opposite case, the parallel case, and the vertical case. These three cases can be formulated independently by using mathematical tools. To verify these formulations, we do lots of simulations by ns2, and the theoretical formulations and simulation results are matched.