2. 南京邮电大学 江苏省无线通信重点实验室, 南京 210003
2. Wireless Communication Key Lab of Jiangsu Province, Nanjing University of Posts and Telecommunications, Nanjing 210003, China
开放科学(资源服务)标志码(OSID):
随着高速多媒体应用和高密度物联网设备的普及,未来5G/B5G网络将会是基站(Base Station,BS)密集部署的热点通信网络,现有的蜂窝网络也将向小区密集化和以用户为中心的小型基站(Small Base Station,SBS)部署方向发展。
由于实际部署在多层5G HetNets中的BS位置表现出不规则性,因此随机几何空间模型被视为HetNets精确建模和分析的重要工具[1-2]。在该方法中,大规模无线网络被抽象为点过程[3-4],从文献[5-6]的研究中可以得出将BS的分布视为一个点过程的原因。通过蒙特卡洛模拟,文献[7]研究得出理想六边形蜂窝系统下行链路信号干扰比的分布,接近根据齐次泊松点过程(Poisson Point Processe,PPP)部署BS的蜂窝系统。在上述文献的启发下,文献[8-10]采用基于随机几何的模型对蜂窝网络性能进行评估,且已有大多数工作都将蜂窝网络建模为传统的理想六边形网格模型。然而,上述研究只将BS和用户设备(User Equipments,UE)在每层中的位置建模为独立的PPP。在基于热点的5G HetNets中,当UE和BS之间存在相关性时,UE和BS之间的独立性假设可能不太准确。在实践中,虽然传统宏基站(Macro Base Station,MBS)的部署较一致,但为了满足全覆盖要求,又部署了其他类型的SBS,如微微基站(Pico Base Station,PBS)和毫微微基站(Femto Base Station,FBS)。因此,SBS有望部署在拥挤或热点地区,以修补覆盖死区,这实际上耦合了UE和SBS的位置,使UE和SBS之间存在一定的相关性[11]。
虽然现有研究大多利用UE和SBS(如PBS)之间的耦合,但在现实中,由于热点区域中密集UE在短时间内引起数据速率突然激增,可能导致以集群为中心的BS过载。在此情况下,需要在以集群为中心的部署模式下部署更多的低功耗FBS作为UE,为过载BS提供流量分流。在实践中,对于高密度HetNets,一个通用的模型是从热点区域中抽象出集群中心并建模为PPP。UE和所有SBS均分散在热点中心周围,并建模为泊松簇过程(Poisson Cluster Process,PCP)。虽然这种部署利用UE与中心之间的耦合以及所有SBS与中心之间的耦合,但对活动节点的位置呈空间分离的场景单纯使用PPP和PCP来建模是不现实的,更合适的模型是泊松洞过程(Poisson Hole Process,PHP)。此外,文献[2-10]仅研究了单层蜂窝网络或者两层异构网络,并未对三层异构网络进行建模和研究。
本文在PPP和PCP的基础上结合PHP,提出一种面向密集热点区域的三层异构网络建模方案,以应用于移动UE和FBS都集中部署或分散在公共集群中心的网络场景。此外,本文还分析有序和非有序FBS这2种情况下不同网络参数对级联概率的影响。
1 网络建模本文考虑一个三层异构蜂窝网络,如图 1所示,其中,第1层、第2层和第3层分别由MBS、PBS和FBS组成,分别称为M层、P层和F层。所有MBS都配有大规模多输入多输出(Multiple-Input Multiple-Output,MIMO)天线
|
Download:
|
| 图 1 三层异构网络系统模型 Fig. 1 System model of three layer heterogeneous network | |
与MBS和PBS的空间分布不同,本文在PHP的辅助下对FBS和UE的空间分布进行建模,PHP是拥挤或热点区域移动UE更真实的空间分布模型。同时,为了进一步提高容量,本文在PBS周围密集部署FBS,从而实现PBS位置与UE和FBS的耦合。此外,除建模UE簇分布,本文还假设剩余的UE服从PHP。
本文将FBS的位置建模为密度
在给定父点过程
由上文可得,目标UE可以与MBS、PBS或FBS相级联,此外,对于给定的级联,目标UE分为簇内UE或簇外UE,所有这些情况使UE类型变得更复杂。图 2所示为UE分类情况,具体如下:
|
Download:
|
| 图 2 UE分类情况 Fig. 2 Classification of UE | |
1)簇中心UE和簇边缘UE过程:簇中心UE过程
2)簇中心FBS和簇边缘FBS过程:簇中心FBS过程
3)簇中心宏小区UE(MUE)、微微小区UE(PUE)和毫微微小区UE(FUE)过程:簇中心MUE、PUE和FUE过程
4)簇边缘MUE和FUE过程:簇边缘MUE和FUE过程
基于簇UE的分类,本文根据带宽分配因子
忽略簇和洞的重叠影响,在随机选取的簇中随机选取一个目标UE,本文将其称为代表簇。根据Slivnyak-Moche定理,在Borel空间上的点过程是PPP,当且仅当Palm分布几乎处处与原分布定理一致。因此,对以
服务于目标UE的BS被称为标记BS。由于MBSs的位置被建模为密度
对于FBSs,本文分别在有序和非有序情况下考虑目标UE到代表簇中FBSs的距离
本文首先描述非有序FBSs情况下的距离
| $ {f}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{\mathrm{N}}‖}\left(a|b\right)=\frac{a}{{\sigma }_{\mathrm{F}}^{2}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{a}^{2}+{b}^{2}}{2{\sigma }_{\mathrm{F}}^{2}}\right){I}_{0}\left(\frac{ab}{{\sigma }_{\mathrm{F}}^{2}}\right) $ | (1) |
其中,
| $ {f}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{\mathrm{N}}‖}\left(x\right)=\frac{x}{\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right) $ | (2) |
在非有序FBSs的情况下,距离
| $ {F}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{\mathrm{N}}‖}\left(x\right)=1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right) $ | (3) |
根据次序统计[17]和类似于式(2)的考虑,在有序FBSs的情况下,目标UE到最近FBS的距离
| $ {F}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{\mathrm{O}}‖}\left(x\right)={\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}} $ | (4) |
根据
| $ \begin{array}{l}{f}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{\mathrm{O}}‖}\left(x\right)={\stackrel{-}{c}}_{\mathrm{F}}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}-1}\mathrm{d}y\cdot \\ \\ \ \ \ \ \ \ \ \mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\frac{x}{\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\end{array} $ | (5) |
在后续分析中,将会删除
由于UEs可能属于簇中心也可能属于簇边缘,因此本文将分别研究簇中心和簇边缘UEs的级联。
考虑到UEs可能有3种类型的级联概率(Association Probability,AP),即MBS、PBS和FBS,以下利用长期平均接收信号功率模型,则目标UE到位于
| $ {P}_{\mathrm{z}, r}^{\mathrm{A}}={G}_{z}\frac{{P}_{z}}{{S}_{z}}L\left(‖{\boldsymbol{x}}_{z}‖\right) $ | (6) |
其中,
在接收信号的平均功率模型式(6)中,首先关注簇内UEs(即簇中心UEs)的级联。由于本文仅考虑蜂窝下行链路传输,因此使用最强的平均偏置接收功率(Average Biased Received Power,ABRP)来决定UEs的级联[19]。ABRP的基本思想是目标UE根据ABRP与强BS相级联,因此,决定服务于目标簇中心UE的BS表示为:
| $ \mathrm{B}{\mathrm{S}}^{\mathrm{C}}:\underbrace {{\mathop{\rm argmax}\nolimits} }_{z \in \left[ {{\rm{M}}, {\rm{P}}, {\rm{F}}} \right]}\left\{{B}_{z}{P}_{z, r}^{\mathrm{A}}\right\} $ | (7) |
其中,
| $ \begin{array}{l}{\widehat{B}}_{kz}={B}_{k}/{B}_{z} , {\widehat{P}}_{kz} ={P}_{k}/{P}_{z}\\ {\widehat{G}}_{kz}=\left({G}_{k}/{S}_{k}\right)/\left({G}_{z}/{S}_{z}\right)\end{array} $ | (8) |
当
命题1 在有序FBSs的情况下,目标簇中心UE与
| $ \begin{array}{l}{\varLambda }_{\mathrm{F}}^{\mathrm{C}\mathrm{O}}= \\ \\ \frac{{\stackrel{-}{c}}_{\mathrm{F}}}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)\right)\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\int }_{0}^{\mathrm{\infty }}x\left(1-\left.\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+ \\ \\ {\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)\right.\right.}^{{\stackrel{-}{c}}_{\mathrm{F}}-1}\left(\stackrel{}{\underset{}{\mathrm{e}\mathrm{x}\mathrm{p}}}\left(\stackrel{}{\underset{}{-}}\left(\stackrel{}{\underset{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}\right.\right.\right.\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}\right.{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\left.{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }+\\ \\ \ \ \ \ \ \ \ \ \frac{1}{2{\sigma }^{2}}{\left({\widehat{B}}_{\mathrm{P}\mathrm{F}} {\widehat{P}}_{\mathrm{P}\mathrm{F}} {\widehat{G}}_{\mathrm{P}\mathrm{F}} \right)}^{2/\alpha }+\left.\left.\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+ \\ \\ {\sigma }_{\mathrm{D}}^{2}\right)}\right){x}^{2}\right)-\mathrm{e}\mathrm{x}\mathrm{p}\left.\left.\left(-\left({\left.\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{x}^{2}\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}\right.{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }+ \\ \\ \frac{1}{2{\sigma }^{2}}{R}_{2}^{2}+ \\ \\ \frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+ \\ \\{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)\right)\mathrm{d}x\right)\end{array} $ | (9) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{M}}^{\mathrm{C}\mathrm{O}}=\frac{2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}\left({\int }_{0}^{\mathrm{\infty }}x\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(\stackrel{}{\underset{}{-}}\right.\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\right.\right.\left({\widehat{B}}_{\mathrm{P}\mathrm{M}}\right.{\widehat{P}}_{\mathrm{P}\mathrm{M}}\left.{\left.{\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}\right)\left.-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right)\cdot \\ \ \ \ \ \ \ \ \ \ \mathrm{e}\mathrm{x}\mathrm{p}\left(-{\rm{ \mathsf{ π} }} {\lambda }_{\mathrm{M}}{x}^{2}\right)\left.\left(1-{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}{\widehat{P}}_{\mathrm{F}\mathrm{M}} {\widehat{G}}_{\mathrm{F}\mathrm{M}} \right)}^{2/\alpha }{x}^{2}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right)\mathrm{d}x\right)\end{array} $ | (10) |
| $ {\varLambda }_{\mathrm{P}}^{\mathrm{C}\mathrm{O}}=\frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)}\left({\int }_{0}^{{R}_{2}}\frac{x}{{\sigma }_{\mathrm{D}}^{2}}\right.\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left.\left({\rm{ \mathsf{ π} }} {\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{P}} {\widehat{P}}_{\mathrm{M}\mathrm{P}} {\widehat{G}}_{\mathrm{M}\mathrm{P}}\right)}^{2/\alpha }{x}^{2}\right)\right)\right.\cdot \\ \\ \left(1-{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{P}} {\widehat{P}}_{\mathrm{F}\mathrm{P}} \\ \\ {\widehat{G}}_{\mathrm{F}\mathrm{P}}\right)}^{2/\alpha }{x}^{2}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\mathrm{d}x\right) $ | (11) |
证明 在有序FBSs的情况下,首先计算目标簇中心UE与位于毫微微小区层
| $ \begin{array}{l}{\varLambda }_{\mathrm{F}}^{\mathrm{C}\mathrm{O}}=\mathbb{P}\left\{{B}_{\mathrm{F}}{P}_{\mathrm{F}, r}>\underset{z\in \left\{\mathrm{M}, \mathrm{P}\right\}}{\mathrm{m}\mathrm{a}\mathrm{x}}\left\{{B}_{\mathrm{z}}{P}_{\mathrm{z}, r}\right\}\right\}=\\ \ \ \ \ \ \ \ \ \ \prod \limits_{z\in \left\{\mathrm{M}, \mathrm{P}\right\}}\mathbb{P}\left\{‖{\boldsymbol{x}}_{\mathrm{z}}‖>{\left({\widehat{B}}_{\mathrm{F}\mathrm{z}}{\widehat{P}}_{\mathrm{F}\mathrm{z}}{\widehat{G}}_{\mathrm{F}\mathrm{z}}\right)}^{1/\alpha }‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖\right\}\end{array} $ | (12) |
在区域
| $ \begin{array}{l} \mathbb{P}\left\{‖{\boldsymbol{x}}_{\mathrm{M}}‖>\right.\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}} \right.\left.{\left.{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{1/\alpha }‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖\right\}=\\ {\mathbb{E}}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{\mathrm{C}}‖}\left\{\mathrm{e}\mathrm{x}\mathrm{p}\right.\left(-\right.\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}\right.\left.\left.{\left.{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }{‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖}^{2}\right)\right\} \end{array} $ | (13) |
注意代表簇中只有一个PBS,并且目标簇中心UEs落在PBS半径为
| $ {F}_{‖{\boldsymbol{x}}_{\mathrm{P}}‖}\left(\boldsymbol{x}\right)=\frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)}\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{‖\boldsymbol{x}‖}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right) $ | (14) |
其中,
| $ \mathbb{P}\left\{‖{\boldsymbol{x}}_{\mathrm{P}}‖>{\left({\widehat{B}}_{\mathrm{P}\mathrm{F}}{\widehat{P}}_{\mathrm{P}\mathrm{F}}{\widehat{G}}_{\mathrm{P}\mathrm{F}}\right)}^{1/\alpha }‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖\right\}= \\ \\ \frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)}{\mathbb{E}}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖}\left\{\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\right.\right.\right.\left({\widehat{B}}_{\mathrm{P}\mathrm{F}}{\widehat{P}}_{\mathrm{P}\mathrm{F}} {\widehat{G}}_{\mathrm{P}\mathrm{F}}\right)\left.\left.\left.{}^{2/\alpha }{‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖}^{2}\right)- \\ \\ \mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right)\right\} $ | (15) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{F}}^{\mathrm{C}\mathrm{O}}=\frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)}\left({\stackrel{}{\underset{}{\mathbb{E}}}}_{‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖}\left\{\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\stackrel{}{\underset{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}\right.\right.\right.\right.\right.\left.\left.\left.{\left.{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }+ \\ \\ \frac{1}{2{\sigma }_{\mathrm{D}}^{2}}{\left({\widehat{B}}_{\mathrm{P}\mathrm{F}}{\widehat{P}}_{\mathrm{P}\mathrm{F}}{\widehat{G}}_{\mathrm{P}\mathrm{F}}\right)}^{2/\alpha }\right){‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖}^{2}\right)\right\}-\\ \ \ \ \ \ \ \ \ \ \left.{\mathbb{E}}_{‖{\boldsymbol{x}}_{F}^{}‖}\left\{\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}} {\widehat{P}}_{\mathrm{M}\mathrm{F}} {G}_{\mathrm{M}}\right)}^{2/\alpha }{‖{\boldsymbol{x}}_{\mathrm{F}}^{}‖}^{2}+\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}{R}_{2}^{2}\right)\right)\right\}\right)\end{array} $ | (16) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{F}}^{\mathrm{C}\mathrm{O}}=\frac{{\stackrel{-}{c}}_{\mathrm{F}}}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)\right)\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\int }_{0}^{\mathrm{\infty }}x \\ \\ {\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}-1}\left(\stackrel{}{\underset{}{\mathrm{e}\mathrm{x}\mathrm{p}}}\left(\stackrel{}{\underset{}{-}}\left({\rm{ \mathsf{ π} }} {\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}} \right)}^{2/\alpha }\right.\right.\right.+\\ \ \ \ \ \ \ \ \ \ \frac{1}{2{\sigma }^{2}}{\left({\widehat{B}}_{\mathrm{P}\mathrm{F}}{\widehat{P}}_{\mathrm{P}\mathrm{F}} {\widehat{G}}_{\mathrm{P}\mathrm{F}} \right)}^{2/\alpha }+\left.\left.\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right){x}^{2}\right)- \\ \\ \mathrm{e}\mathrm{x}\mathrm{p}\left.\left(-\left({\left.{\rm{ \mathsf{ π} }} {\lambda }_{\mathrm{M}}{x}^{2}\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}\right.{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }+\frac{1}{2{\sigma }^{2}}{R}_{2}^{2}+\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)\right)\mathrm{d}x\end{array} $ | (17) |
当目标簇中心UE与式(12)的具有对称性的MBS相级联时,级联概率
| $ {\varLambda }_{\mathrm{M}}^{\mathrm{C}\mathrm{O}}=\prod \limits_{z\in \left\{\mathrm{P}, \mathrm{F}\right\}}\mathbb{P}\left\{‖{\boldsymbol{x}}_{\mathrm{z}}‖>{\left({\widehat{B}}_{zM}{\widehat{P}}_{zM} {\widehat{G}}_{zM}\right)}^{1/\alpha }‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖\right\} $ | (18) |
当
| $ \mathbb{P}\left\{‖{\boldsymbol{x}}_{\mathrm{P}}‖>{\left({\widehat{B}}_{\mathrm{P}\mathrm{M}}{\widehat{P}}_{\mathrm{P}\mathrm{M}} {\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{1/\alpha }‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖\right\}= \\ \\ \frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}{\mathbb{E}}_{‖{\bf{x}}_{\mathrm{M}}^{}‖} \\ \\ \left\{\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\times {\left({\widehat{B}}_{\mathrm{P}\mathrm{M}}{\widehat{P}}_{\mathrm{P}\mathrm{M}}{\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }{‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖}^{2}\right)\right.\right.\left.\left. \\ \\ -\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right)\right\} $ | (19) |
| $ \mathbb{P}\left\{‖{\boldsymbol{x}}_{\mathrm{F}}‖>{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}{\widehat{P}}_{\mathrm{F}\mathrm{M}}{\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{1/\alpha }‖{\boldsymbol{x}}_{\mathrm{M}}‖\right\}= \\ \\ 1-{\mathbb{E}}_{‖{\boldsymbol{x}}_{\mathrm{M}}^{\mathrm{C}}‖}\left\{\left.\left(1-\mathrm{e}\mathrm{x}\mathrm{p}{\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+ \\ \\ {\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}{\widehat{P}}_{\mathrm{F}\mathrm{M}}{\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }{‖{\boldsymbol{x}}_{\mathrm{M}}‖}^{2}\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right)\right\}\right. $ | (20) |
| $ \begin{array}{l} {\varLambda }_{\mathrm{M}}^{\mathrm{C}\mathrm{O}}= \\ \\ \frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}\left({\mathbb{E}}_{‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖}\left\{\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}{\left({\widehat{B}}_{\mathrm{P}\mathrm{M}}{\widehat{P}}_{\mathrm{P}\mathrm{M}} {\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }{‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖}^{2}\right)\right.\right.\right. \\ \\ -\mathrm{e}\mathrm{x}\mathrm{p}\left.\left.\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right)\right\}-\\ \ \ \ \ \ \ \ \ {\mathbb{E}}_{‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖}\left\{\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\left({\widehat{B}}_{\mathrm{P}\mathrm{M}} {\widehat{P}}_{\mathrm{P}\mathrm{M}}\right.\right.\right.\right.\left.{\left.{\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }{‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖}^{2}\right)\left. \\ \\ -\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right)\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right.\right.\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}\right. \\ \\ {\left.{\widehat{P}}_{\mathrm{F}\mathrm{M}}{\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }\left.\left.{\left.\left.{‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖}^{2}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right\}\right) \end{array} $ | (21) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{M}}^{\mathrm{C}\mathrm{O}}=\frac{2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}{\int }_{0}^{\mathrm{\infty }}x\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\left({\widehat{B}}_{\mathrm{P}\mathrm{M}}{\widehat{P}}_{\mathrm{P}\mathrm{M}}\right.\right.\right.\left.{\left.{\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}\right)-\mathrm{e}\mathrm{x}\mathrm{p}\left.\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right)\mathrm{e}\mathrm{x}\mathrm{p}\left(-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{x}^{2}\right)\cdot \\ \ \ \ \ \ \ \ \ \ \ \left(1-{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}} {\widehat{P}}_{\mathrm{F}\mathrm{M}}{\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right)\mathrm{d}x\end{array} $ | (22) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{P}}^{\mathrm{C}\mathrm{O}}=\frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(\frac{-{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)}{\int }_{0}^{{R}_{2}}\frac{x}{{\sigma }_{\mathrm{D}}^{2}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}\right.\right.\left({\widehat{B}}_{\mathrm{M}\mathrm{P}}\right.\left.\left.{\left.{\widehat{P}}_{\mathrm{M}\mathrm{P}} {\widehat{G}}_{\mathrm{M}\mathrm{P}}\right)}^{2/\alpha }{x}^{2}\right)\right)\cdot \\ \ \ \ \ \ \ \ \ \left({\left(1-\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{P}}{\widehat{P}}_{\mathrm{F}\mathrm{P}}{\widehat{G}}_{\mathrm{F}\mathrm{P}}\right)}^{2/\alpha }{x}^{2}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right)\mathrm{d}x\end{array} $ | (23) |
利用与命题1相似的论据,可以得到在非有序FBSs情况下的AP,即推论1,其中推论1可以用与命题1相似的准则来证明。
推论1 在非有序FBSs的情况下,目标簇中心UEs与
| $ \begin{array}{l}{\varLambda }_{\mathrm{F}}^{\mathrm{C}\mathrm{N}}=\frac{1}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)\right)\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}x{\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\right.\left(\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}} \right)}^{2/\alpha }+\frac{1}{2{\sigma }^{2}}\left({\widehat{B}}_{\mathrm{P}\mathrm{F}}{\widehat{P}}_{\mathrm{P}\mathrm{F}}\right.\right.\left. {\widehat{G}}_{\mathrm{P}\mathrm{F}}\right)\right.}^{2/\alpha }+\\ \left.\\ \ \ \ \ \ \ \ \ \ \left.\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right){x}^{2}\right)-\mathrm{e}\mathrm{x}\mathrm{p}\left.\left.\left(-\left({\stackrel{\stackrel{}{}}{\underset{}{\mathrm{{\rm{ \mathsf{ π} }} }\lambda }}}_{\mathrm{M}}{x}^{2}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }\right.+\frac{1}{2{\sigma }^{2}}{R}_{2}^{2}+\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)\right)\mathrm{d}x\end{array} $ | (24) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{M}}^{\mathrm{C}\mathrm{N}}=\frac{2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}{\int }_{0}^{\mathrm{\infty }}x \left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\right.\right.\right.\left({\widehat{B}}_{\mathrm{P}\mathrm{M}}\right.{\left.{\widehat{P}}_{\mathrm{P}\mathrm{M}} {\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }+ \\ \\ \frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+ \\ \\ {\sigma }_{\mathrm{D}}^{2}\right)}\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}\right.{\widehat{P}}_{\mathrm{F}\mathrm{M}}{\left. {\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }+\left.\left.\stackrel{\stackrel{}{}}{\underset{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}\right){x}^{2}\right)-\\ \ \ \ \ \ \ \ \ \ \ \ \mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}{R}_{2}^{2}+\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right.\right.\left.\left.\left. \\ \\ {\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}{\widehat{P}}_{\mathrm{F}\mathrm{M}} {\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}+\stackrel{}{\underset{}{\stackrel{}{\underset{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}}}{x}^{2}\right)\right)\right)\mathrm{d}x\end{array} $ | (25) |
| $ {\varLambda }_{\mathrm{P}}^{\mathrm{C}\mathrm{N}}=\frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}{\int }_{0}^{{R}_{2}^{}}\frac{x}{{\sigma }_{\mathrm{D}}^{2}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\stackrel{}{\underset{}{\stackrel{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}}\right.\left(\stackrel{}{\underset{}{\left({\widehat{B}}_{\mathrm{M}\mathrm{P}}{\widehat{P}}_{\mathrm{M}\mathrm{P}}\right.}}\right.{\left.{\widehat{G}}_{\mathrm{M}\mathrm{P}}\right)}^{2/\alpha }+\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{P}} {\widehat{P}}_{\mathrm{F}\mathrm{P}} {\widehat{G}}_{\mathrm{F}\mathrm{P}}\right)}^{2/\alpha }+ \\ \\ \left.\left.\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\right){x}^{2}\right)\mathrm{d}x $ | (26) |
由命题1易得目标UE到z层服务BS级联距离的PDF。在有序FBSs的情况下,假设服务FBS、MBS和PBS分别位于
命题2 在有序FBSs的情况下,假设目标簇中心UE与
| $ \begin{array}{l}{f}_{{X}_{\mathrm{F}}^{\mathrm{C}\mathrm{O}}}\left(\boldsymbol{x}\right)=\frac{1}{{\varLambda }_{\mathrm{F}}^{\mathrm{C}\mathrm{O}}}\frac{{\stackrel{-}{c}}_{\mathrm{F}}x}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)\right)\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}-1}\cdot \\ \ \ \ \ \ \ \ \ \mathrm{e}\mathrm{x}\mathrm{p}\left(-{x}^{2}\left(\stackrel{}{\underset{}{\stackrel{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}}\right.\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)\right.\left.\left.{}^{2/\alpha }+\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }^{2}}\right.\times \right.\left.\left.{\left({\widehat{B}}_{\mathrm{P}\mathrm{F}} {\widehat{P}}_{\mathrm{P}\mathrm{F}} {\widehat{G}}_{\mathrm{P}\mathrm{F}} \right)}^{2/\alpha }\stackrel{}{\underset{}{\stackrel{}{{x}^{2}}}}\right)- \\ \\ \mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }^{2}}{R}_{2}^{2}\right)\right)\end{array} $ | (27) |
| $ \begin{array}{l}{f}_{{X}_{\mathrm{M}}^{\mathrm{C}\mathrm{O}}}\left(\boldsymbol{x}\right)=\frac{1}{{\varLambda }_{\mathrm{M}}^{\mathrm{C}\mathrm{O}}}\frac{2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}x}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{x}^{2}\right)\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}{\left({\widehat{B}}_{\mathrm{P}\mathrm{M}}{\widehat{P}}_{\mathrm{P}\mathrm{M}}{\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }\right.\right.\left.{\stackrel{}{\underset{}{‖{\boldsymbol{x}}_{\mathrm{M}}^{}‖}}}^{2}\right)- \\ \ \ \ \ \ \ \ \ \left.\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\right)\left(1-{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}{\widehat{P}}_{\mathrm{F}\mathrm{M}}{\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right) \end{array} $ | (28) |
| $ \begin{array}{l}{f}_{{X}_{\mathrm{P}}^{\mathrm{C}\mathrm{O}}}\left(\boldsymbol{x}\right)=\frac{1}{{\varLambda }_{\mathrm{P}}^{\mathrm{C}\mathrm{O}}}\frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)}\frac{x}{{\sigma }_{\mathrm{D}}^{2}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{x}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{P}}{\widehat{P}}_{\mathrm{M}\mathrm{P}}{\widehat{G}}_{\mathrm{M}\mathrm{P}}\right)}^{2/\alpha }{x}^{2}\right)\right)\cdot \\ \ \ \ \ \ \ \ \ \left(1-{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{P}} {\widehat{P}}_{\mathrm{F}\mathrm{P}} {\widehat{G}}_{\mathrm{F}\mathrm{P}}\right)}^{2/\alpha }{x}^{2}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right)\end{array} $ | (29) |
此外,在非有序FBSs的情况下,推论2给出了相应级联距离的PDF。
推论2 在非有序FBSs的情况下,假设目标簇中心UE与
| $ \begin{array}{l} {f}_{{X}_{\mathrm{M}}^{\mathrm{C}\mathrm{N}}}\left(\boldsymbol{x}\right)=\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}{\left({\widehat{B}}_{\mathrm{P}\mathrm{M}} {\widehat{P}}_{\mathrm{P}\mathrm{M}} {\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }+\right.\right.\left.\left.\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}} {\widehat{P}}_{\mathrm{F}\mathrm{M}} {\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }+\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}\right){x}^{2}\right)- \\ \ \ \ \ \ \ \ \ \frac{1}{{\varLambda }_{\mathrm{M}}^{\mathrm{C}\mathrm{N}}}\frac{2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}x}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}{R}_{2}^{2}+\right.\right.\left.\left.\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}{\widehat{P}}_{\mathrm{F}\mathrm{M}} {\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}+\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{x}^{2}\right)\right)\end{array} $ | (30) |
| $ \begin{array}{l}{f}_{{X}_{\mathrm{F}}^{\mathrm{C}\mathrm{N}}}\left(\boldsymbol{x}\right)= \\ \\ \frac{1}{{\varLambda }_{\mathrm{F}}^{\mathrm{C}\mathrm{N}}}\frac{x}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-{R}_{2}^{2}/2{\sigma }_{\mathrm{D}}^{2}\right)\right)\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\left(\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}} {\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }+ \\ \\ \frac{1}{2{\sigma }^{2}}\right.\right.\right.\left({\widehat{B}}_{\mathrm{P}\mathrm{F}}\right.{\left.{\widehat{P}}_{\mathrm{P}\mathrm{F}}{\widehat{G}}_{\mathrm{P}\mathrm{F}}\right)}^{2/\alpha }+\left.\left.\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right) {x}^{2}\right) -\\ \ \ \ \ \ \ \ \ \\ \\ \mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\stackrel{}{\underset{}{\stackrel{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }\right.+\frac{1}{2{\sigma }^{2}}\right.{R}_{2}^{2}\left.\left.\left.+\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+ \\ \\ {\sigma }_{\mathrm{D}}^{2}\right)}\right){x}^{2}\right)\right)\end{array} $ | (31) |
| $ {f}_{{X}_{\mathrm{P}}^{\mathrm{C}\mathrm{N}}}\left(\boldsymbol{x}\right)= \\ \\ \frac{1}{{\varLambda }_{\mathrm{P}}^{\mathrm{C}\mathrm{N}}}\frac{1}{1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{R}_{2}^{2}}{2{\sigma }_{\mathrm{D}}^{2}}\right)}\frac{x}{{\sigma }_{\mathrm{D}}^{2}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{P}}{\widehat{P}}_{\mathrm{M}\mathrm{P}}{\widehat{G}}_{\mathrm{M}\mathrm{P}}\right)}^{2/\alpha }+\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+ \\ \\ {\sigma }_{\mathrm{D}}^{2}\right)}\right.\right.\left.\left.{\left({\widehat{B}}_{\mathrm{F}\mathrm{P}}{\widehat{P}}_{\mathrm{F}\mathrm{P}}{\widehat{G}}_{\mathrm{F}\mathrm{P}}\right)}^{2/\alpha }+\frac{1}{2{\sigma }_{\mathrm{D}}^{2}}\right){x}^{2}\right) $ | (32) |
在目标UE位于PBSs的覆盖范围之外时,簇边缘UEs仅具有两种可能的级联类型,即MBS和FBS,如图 1所示。级联准则式(7)可以重新记为:
| $ \mathrm{B}{\mathrm{S}}^{\mathrm{E}}:\underbrace {{\mathop{\rm argmax}\nolimits} }_{z \in \left[ {{\rm{M}}, {\rm{F}}} \right]} \left\{{B}_{z}{P}_{z, r}^{A}\right\} $ | (33) |
因此,对于有序FBSs的情况,可以得到命题3。
命题3 在有序FBSs的情况下,目标簇边缘UE与
| $ \begin{array}{l}{\varLambda }_{\mathrm{F}}^{\mathrm{E}\mathrm{O}}={\int }_{0}^{\mathrm{\infty }}\frac{{\stackrel{-}{c}}_{\mathrm{F}}x}{\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{{a}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}-1}\cdot \\ \ \ \ \ \ \ \ \ \ \mathrm{e}\mathrm{x}\mathrm{p}\left(-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }{x}^{2}-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\mathrm{d}x\end{array} $ | (34) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{M}}^{\mathrm{E}\mathrm{O}}=1-2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\int }_{0}^{\mathrm{\infty }}x\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right.\right.\cdot \\ \ \ \ \ \ \ \ \ \ {\left.\left.\left({\left({\widehat{B}}_{\mathrm{F}\mathrm{M}} {\widehat{P}}_{\mathrm{F}\mathrm{M}} {\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}\right)\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{x}^{2}\right)\mathrm{d}x\end{array} $ | (35) |
推论3 在非有序FBSs的情况下,目标簇边缘UE与
| $ \begin{array}{l}{\varLambda }_{\mathrm{F}}^{\mathrm{E}\mathrm{N}}={\int }_{0}^{\mathrm{\infty }}\frac{x}{\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\mathrm{e}\mathrm{x}\mathrm{p}\left(\stackrel{\stackrel{}{}}{\underset{}{-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}\right.\right.{\left.{\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }{x}^{2}-\\ \ \ \ \ \ \ \ \ \ \left.\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\mathrm{d}x\end{array} $ | (36) |
| $ \begin{array}{l}{\varLambda }_{\mathrm{M}}^{\mathrm{E}\mathrm{N}}=2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{M}{\int }_{0}^{\mathrm{\infty }}x\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\frac{1}{2\left({\sigma }_{\mathrm{D}}^{2}+{\sigma }_{\mathrm{F}}^{2}\right)}\right.\right.{\left({\widehat{B}}_{\mathrm{F}\mathrm{M}}{\widehat{P}}_{\mathrm{F}\mathrm{M}}{\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }+\\ \ \ \ \ \ \ \ \ \left.\left.\stackrel{}{\underset{}{\stackrel{}{\underset{}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}}}\right){x}^{2}\right)\mathrm{d}x\end{array} $ | (37) |
此外,在有序FBSs的情况下,目标簇边缘UE与其服务FBS和MBS的级联距离分别为
命题4 在有序FBSs的情况下,假设目标簇边缘UE与
| $ \begin{array}{l}{f}_{{X}_{\mathrm{F}}^{\mathrm{E}\mathrm{O}}}\left(\boldsymbol{x}\right)=\frac{1}{{\varLambda }_{\mathrm{F}}^{\mathrm{E}\mathrm{O}}}\frac{{\stackrel{-}{c}}_{\mathrm{F}}x}{\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}{\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(\frac{-{a}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}-1}\cdot \\ \ \ \ \ \ \ \ \ \ \mathrm{e}\mathrm{x}\mathrm{p}\left(-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}} {\widehat{P}}_{\mathrm{M}\mathrm{F}}{\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }{x}^{2}-\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\end{array} $ | (38) |
| $ \begin{array}{l}{f}_{{X}_{\mathrm{M}}^{\mathrm{E}\mathrm{O}}}\left(\boldsymbol{x}\right)=\frac{2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}x}{{\varLambda }_{\mathrm{M}}^{\mathrm{E}\mathrm{O}}}\left(1-\left(1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{1}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right.\right.\ \cdot \right.\\ \ \ \ \ \ \ \ \ \ \ \ \left.{\left.\left.\left({\left({\widehat{B}}_{\mathrm{F}\mathrm{M}} {\widehat{P}}_{\mathrm{F}\mathrm{M}} {\widehat{G}}_{\mathrm{F}\mathrm{M}}\right)}^{2/\alpha }{x}^{2}\right)\right)\right)}^{{\stackrel{-}{c}}_{\mathrm{F}}}\right)\mathrm{e}\mathrm{x}\mathrm{p}\left(-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}{x}^{2}\right)\end{array} $ | (39) |
此外,在非有序FBSs的情况下,目标簇边缘UE与其服务FBS和MBS的级联距离分别为
推论4 在非有序FBSs的情况下,假设目标簇边缘UE与
| $ \begin{array}{l}{f}_{{X}_{\mathrm{F}}^{\mathrm{E}\mathrm{N}}}\left(\boldsymbol{x}\right)=\frac{1}{{\varLambda }_{\mathrm{F}}^{\mathrm{E}\mathrm{N}}}\frac{x}{\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\mathrm{e}\mathrm{x}\mathrm{p}\left(\stackrel{\stackrel{}{}}{\underset{}{-\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}\right.{\left({\widehat{B}}_{\mathrm{M}\mathrm{F}}{\widehat{P}}_{\mathrm{M}\mathrm{F}} {\widehat{G}}_{\mathrm{M}\mathrm{F}}\right)}^{2/\alpha }{x}^{2}-\\ \ \ \ \ \ \ \ \ \ \ \ \left.\frac{{x}^{2}}{2\left({\sigma }_{\mathrm{F}}^{2}+{\sigma }_{\mathrm{D}}^{2}\right)}\right)\end{array} $ | (40) |
| $ \begin{array}{l}{f}_{{X}_{\mathrm{M}}^{\mathrm{E}\mathrm{N}}}\left(\boldsymbol{x}\right)=\frac{2\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}x}{{\varLambda }_{\mathrm{M}}^{\mathrm{E}\mathrm{N}}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-\left(\frac{1}{2\left({\sigma }_{\mathrm{D}}^{2}+{\sigma }_{\mathrm{F}}^{2}\right)}\right.\right.{\left({\widehat{B}}_{\mathrm{P}\mathrm{M}} {\widehat{P}}_{\mathrm{P}\mathrm{M}}{\widehat{G}}_{\mathrm{P}\mathrm{M}}\right)}^{2/\alpha }+\\ \ \ \ \ \ \ \ \ \ \ \ \ \ \left.\left.\stackrel{}{\underset{\underset{}{}}{\mathrm{{\rm{ \mathsf{ π} }} }{\lambda }_{\mathrm{M}}}}\right)\stackrel{}{{x}^{2}}\right)\end{array} $ | (41) |
本章节进行模拟实验和数值结果分析,以验证HetNets建模方案的正确性,并说明不同网络参数对级联概率的影响。在整个网络中,假设所有链路的路径损耗指数均为
|
下载CSV 表 1 系统参数及其取值 Table 1 System parameters and their values |
图 3所示为方差
|
Download:
|
| 图 3 目标簇中心UE与FBS、MBS和PBS的级联概率 Fig. 3 Association probability of target cluster-center UE with FBS, MBS and PBS | |
|
Download:
|
| 图 4 目标簇边缘UE与FBS和MBS的级联概率 Fig. 4 Association probability of target cluster-edge UE with FBS and MBS | |
从图 3可以看出,分别以方差
由于只有MUEs和FUEs不在PBS的覆盖范围之内,因此图 4给出了
本文针对密集热点通信场景,提出一种三层HetNets建模方案。分析不同网络参数对UEs级联概率的影响,并在有序FBSs和非有序FBSs两种情况下,分别对比目标UE与FBS、MBS和PBS级联的AP大小。下一步将在本文方案的基础上对网络覆盖概率进行研究,并且将在该网络模型中加入D2D用户,研究D2D通信模式下的网络覆盖概率。
| [1] |
CHIU S N, STOYAN D, KENDALL W S, et al. Stochastic geometry and its applications[M]. Washington D.C., USA: John Wiley & Sons, 2013.
|
| [2] |
ANDREWS J G, BACCELLI F, GANTI R K. A tractable approach to coverage and rate in cellular networks[J]. IEEE Transactions on Communications, 2011, 59(11): 3122-3134. DOI:10.1109/TCOMM.2011.100411.100541 |
| [3] |
HASAN A, ANDREWS J G. The guard zone in wireless ad hoc networks[J]. IEEE Transactions on Wireless Communications, 2007, 6(3): 897-906. DOI:10.1109/TWC.2007.04793 |
| [4] |
GANTI R K, HAENGGI M. Interference and outage in clustered wireless ad hoc networks[J]. IEEE Transactions on Information Theory, 2009, 55(9): 4067-4086. DOI:10.1109/TIT.2009.2025543 |
| [5] |
LEE C, HAENGGI M. Interference and outage in Poisson cognitive networks[J]. IEEE Transactions on Wireless Communications, 2012, 11(4): 1392-1401. DOI:10.1109/TWC.2012.021512.110131 |
| [6] |
CHEN C, ELLIOTT R C, KRZYMIEN W A. Downlink coverage analysis of n-tier heterogeneous cellular networks based on clustered stochastic geometry[C]//Proceedings of 2013 Asilomar Conference on Signals, Systems and Computers. Washington D.C., USA: IEEE Press, 2013: 1577-1581.
|
| [7] |
BROWN T X. Practical cellular performance bounds via shotgun cellular systems[J]. IEEE Journal on Selected Areas in Communications, 2000, 18(11): 2443-2455. DOI:10.1109/49.895048 |
| [8] |
XIE B, ZHANG Z, HU R Q, et al. Joint spectral efficiency and energy efficiency in FFR-based wireless heterogeneous networks[J]. IEEE Transactions on Vehicular Technology, 2018, 67(9): 8154-8168. DOI:10.1109/TVT.2017.2701356 |
| [9] |
HE A, WANG L, YUE C, et al. Throughput and energy efficiency for S-FFR in massive MIMO enabled heterogeneous C-RAN[C]//Proceedings of GLOBECOM'16. Washington D.C., USA: IEEE Press, 2016: 1-6.
|
| [10] |
ALTAY C, KOCA M. Fractional frequency reuse in non-orthogonal multiple access heterogeneous networks[C]//Proceedings of 2018 IEEE International Conference on Communications. Washington D.C., USA: IEEE Press, 2018: 1-6.
|
| [11] |
SAHA C, AFSHANG M, DHILLON H S. Poisson cluster process: bridging the gap between PPP and 3GPP HetNet models[C]//Proceedings of 2017 Information Theory and Applications Workshop. Washington D.C., USA: IEEE Press, 2017: 1-9.
|
| [12] |
CHUN Y J, HASNA M O, GHRAYEB A. Modeling heterogeneous cellular networks interference using Poisson cluster processes[J]. IEEE Journal on Selected Areas in Communications, 2015, 33(10): 2182-2195. DOI:10.1109/JSAC.2015.2435271 |
| [13] |
AFSHANG M, DHILLON H S. Poisson cluster process based analysis of HetNets with correlated user and base station locations[J]. IEEE Transactions on Wireless Communications, 2018, 17(4): 2417-2431. DOI:10.1109/TWC.2018.2794983 |
| [14] |
YI W, LIU Y, NALLANATHAN A. Modeling and analysis of D2D millimeter-wave networks with poisson cluster processes[J]. IEEE Transactions on Communications, 2017, 65(12): 5574-5588. DOI:10.1109/TCOMM.2017.2744644 |
| [15] |
TURGUT E, GURSOY M C. Uplink performance analysis in D2D-enabled millimeter-wave cellular networks with clustered users[J]. IEEE Transactions on Wireless Communications, 2019, 18(2): 1085-1100. DOI:10.1109/TWC.2018.2889755 |
| [16] |
AFSHANG M, DHILLON H S, CHONG P H J. Modeling and performance analysis of clustered device-to-device networks[J]. IEEE Transactions on Wireless Communications, 2016, 15(7): 4957-4972. |
| [17] |
DAVID H A, NAGARAJA H N. Order statistics[M]. New York, USA: John Wiley and Sons, 1970.
|
| [18] |
WANG L, WONG K K, ELKASHLAN M, et al. Secrecy and energy efficiency in massive MIMO aided heterogeneous C-RAN: a new look at interference[J]. IEEE Journal of Selected Topics in Signal Processing, 2016, 10(8): 1375-1389. DOI:10.1109/JSTSP.2016.2600520 |
| [19] |
SINGH S, ZHANG X, ANDREWS J G. Joint rate and SINR coverage analysis for decoupled uplink-downlink biased cell associations in HetNets[J]. IEEE Transactions on Wireless Communications, 2015, 14(10): 5360-5373. DOI:10.1109/TWC.2015.2437378 |
| [20] |
JIA X, JI P, CHEN Y. Modeling and analysis of multi-tier clustered millimeter-wave cellular networks with user classification for large-scale hotspot area[J]. IEEE Access, 2019, 7: 140278-140299. DOI:10.1109/ACCESS.2019.2943687 |
2021, Vol. 47
