购物中心传播场景包括一个开放的空间,周围的小房间,例如商店。开放空间通常包含了一些障碍,比如餐饮领域,自动扶梯。此外,可能会有多个层商店在不同的楼层。室内问题高度的范围被认为是0.5每层3米,0.5 -15在不同的楼层和之间的距离问题外墙面米或更长时间根据商场的大小。在开放空间,它假定有反映归纳相当长的延迟相比,室内封闭的空间。问题的运动比较缓慢,从0公里/小时的速度变化(餐饮面积和内部店)到10公里/小时(步行空间和画廊)。额外的损失造成的人体跟踪时可能需要考虑拥挤的购物中心。身体影响,然而,失踪在最初的信道模型。通过混凝土墙渗透和衍射损失以及玻璃窗户/门可能需要考虑,当室内和室外之间的干扰用户必须考虑。
Urban road vehicle-to-vehicle (V2V) 城市道路车辆到车辆(V2V)
In the urban V2 V scenario, the road consists of two to four lanes, the houses are closer to the curb, and the traffic density is higher than the other scenarios. The V2 V distance is usually smaller than 300 m and the vehicle antenna height is in the range of 1–3 m. In the urban road, the velocity of vehicles is about 0–60 km/h because of the high traffic density [19]. The propagation includes the penetration caused by the cars, buildings and other obstacles along the roadside [20, 21].
在城市V2V场景中,道路由两个四车道,房子靠近路边,交通密度高于其他场景。V2 V通常小于300米,距离和车辆天线高度是1 - 3 m的范围。在城市道路,车辆的速度大约是0-60 km / h因为交通密度高[19]。传播包括渗透造成的汽车,沿着路边建筑物和其他障碍(20、21)。
Highway V2V 高速公路V2V
In the highway V2 V scenario, there are two to four lanes in each direction with few buildings and cars around, thus the velocity is much higher than that in the urban road. The V2 V distance is normally larger than 1 km or ranges from 300 m to 1 km, and the vehicle antenna height is in the range of 1–3 m. In the highway,the velocity of vehicles is 60–120 km/h, which is higher than the road traffic scenario. The propagation is mainly caused by the natural environment and metal fences on the highway side [22].
在高速公路V2 V的场景中,有两到四车道在每个方向上很少有建筑和汽车,因此速度远远高于城市道路。V2 V通常大于1公里距离或范围从300到1公里,和汽车天线高度是1 - 3 m的范围。在高速公路上,车辆的速度是60 - 120公里/小时,这是高于道路交通场景。造成的传播主要是自然环境和金属栅栏在高速公路上[22]。
Based on the aforementioned scenarios, several D2D channel measurement campaigns have been built up recently. These measurement campaigns only cover parts of the above-defined scenarios and more measurements are needed for better understanding of D2D channels. Compared with the conventional cellular link, the major characteristics of D2D channels can be observed in the following two aspects: (i) both the transmitter and the receiver have low elevation antennas in D2D links, whereas either the transmitter or the receiver is located at higher grounds (above rooftops/ceiling/close to rooftops) in the traditional cellular link; (ii) both the transmitter and the receiver could be moving in D2D links (dual mobility), while at most one is moving in a traditional cellular link. These factors contribute significantly to the specific channel propagation characteristics of D2D channels [23].
基于上述场景,几个D2D信道测量运动最近建立了。这些测量运动只覆盖部分上述定义场景和更多的测量需要更好地理解D2D频道。相比与传统的蜂窝连接,D2D渠道的主要特点可以观察到在以下两个方面:(i)发射机和接收机低海拔天线在D2D链接,而发射机或接收机位于更高的理由(上图屋顶天花板/接近屋顶)在传统的蜂窝联系;(2)发射机和接收机可以朝着D2D链接(双移动),而一个正朝着一个传统的蜂窝连接。这些因素显著的特定信道传播特性D2D频道[23]。
Symmetric and low UE antenna height: D2D path loss should reflect the fact that the nodes have nearly the same low antenna height, thus higher signal attenuation can be expected for the same distance between UEs and the probability of LoS against distance is expected to be lowered. The scattering environment on both sides of transmission link has similar statistical properties, such as the distributions of angle of departure (AoD) and angle of arrival(AoA).
对称天线高度和低问题:D2D路径损耗应该反映这一事实节点几乎相同的天线高度较低,因此可以预期更高的信号衰减相同的距离问题和洛杉矶的概率距离预计将降低。散射环境传播双方的链接也有类似的统计特性,如离去角的分布(AoD)和到达角(AoA)。
UE density/proximity: Nearby UEs have high probability of LoS propagation, thus stochastic system level analysis requires the introduction of LoS probability or some break point distance.Interlink dependency is expected to be higher owing to UE proximity. Effect of shadow fading correlation for D2D links and cellular links will be more noticeable.
问题密度/距离:附近的议题LoS传播的概率很高,因此随机系统级分析要求引入洛杉矶概率或一些断点的距离。连接依赖有望更高由于问题接近。影响D2D阴影衰落相关性的链接和蜂窝链接会更明显。
Mobility: Terminals and surrounding objects are moving, which results in large Doppler spread compared with cellular links, where the BS is static and the variation of large-scale channel propagation parameters is over long periods of time.
流动性:终端和周围的物体移动,导致大的多普勒扩散与蜂窝连接相比,b是静态和大规模的信道传播参数的变化在很长一段时间。
All the aforementioned factors have impact on both D2D small-scale and large-scale channel propagation characteristics and should be observed and discussed by channel measurements.Currently, most D2D channel measurements have concentrated on the discovery of the following significant D2D channel characteristics, as shown in Table 2.
所有上述因素影响D2D小规模和大规模的信道传播特性,应该观察和讨论信道测量。目前,大多数D2D信道测量集中在以下重要的发现D2D信道特性,如表2所示。
Large-scale parameter 大数据
The path loss and shadow fading are the main large-scale channel factors that have major impact on D2D channel modelling. One of the WINNER II models [24] is from indoor office measurement,where the access point is located in corridor, and the UE is randomly dropped within the indoor office. The antenna height for both nodes ranges from 1 to 2.5 m. Besides, 3GPP indoor pico model [25] is applicable for indoor hotspot evaluation. This model assumes the access point is located on the ceiling, and the UE is uniformly dropped within the indoor area. One shortcoming is that the antenna is modelled as being installed on the ceiling. These models enable performance comparisons between D2D and other technical solutions and ensure a consistency in simulation results.
路径损耗和阴影衰落是主要的大型渠道因素D2D频道造型产生重大影响。获胜者之一二
世从室内办公室测量模型[24],访问点位于走廊,室内办公室内的随机下降问题。两个节点的天线高度范围从1到2.5。此外,3gpp室内pico模型[25]适用于室内热点评估。该模型假设访问点位于天花板,问题是在室内面积均匀下降。一个缺点是天线建模是安装在天花板上。这些模型支持D2D之间的性能比较和其他技术解决方案和保证仿真结果的一致性。 LoS/non-LoS (NLoS)
Most measurement models include LoS and NLoS models. For NLoS, WINNER II model [24] introduces penetrations of indoor–outdoor propagation, such as light walls, heavy walls and floor loss. So the total path loss depends on the type of link (LoS or NLoS), the type of walls, the number of walls and the number of floors. The propagation model [25] includes the LoS and NLoS components depending on the distance between nodes. Owing to the low antenna heights and I2I scenarios, its formulas for path loss calculation are simple and easy to use for D2D communication evaluation, and the measurements from these LoS and NLoS models are suitable for D2D channel modelling.
大多数测量模型包括洛杉矶和仿真结果模型。对于仿真结果,赢家II模型[24]介绍了室内外传播、渗透等光墙,沉重的墙壁和地板上的损失。所以总路径损耗取决于类型的链接(洛杉矶或仿真结果),类型的墙壁,墙壁和地板的数量。传播模型[25]包括洛杉矶和仿真结果组件根据节点之间的距离。由于较低的天线高度和I2I场景,其路径损耗的公式计算简单和易于使用的D2D沟通评估,这些洛的测量和仿真结果模型是适合D2D信道模型。
K-factor K-因子
For the ITU UMi and InH channel models [25], the Ricean K-factor (in dB) is modelled as a Gaussian random variable with fixed mean and standard deviation values for LoS. It should be noted that, while most works reported in the literature confirm the choice of a fixed mean K-factor, there are some works (e.g. [26, 27]) that propose linear distant-dependent models for the mean K-factor for LoS scenarios. Unfortunately, no other measurement-based results are available in the literature to validate the sensitivity suggested by the model in [26]. Local scatterers around the transmitter and the receiver (because of low elevation antennas) in D2D channel create different K-factor characteristics compared with traditional communication links.
ITU UMi和异烟肼信道模型[25],Ricean的增殖系数(dB)建模为高斯随机变量与洛杉矶的固定的平均值和标准偏差值。应该注意的是,虽然大多数工作报告在文献中确定一个固定的选择意味着增殖系数,有一些作品(例如(26、27)),提出线性distant-dependent模型为洛杉矶的情况意味着增殖系数。不幸的是,没有其他文献中出现的计量结果来验证提出的灵敏度模型[26]。当地散射在发射机和接收机(因为低海拔的天线)在D2D信道创建不同的增殖系数特征与传统的通信链路。
Delay spread 延迟扩散
For the power delay profile (PDP), the most relevant prior works in the literature have adopted single-slope exponential PDP, including the 3GPP SCM models [28]. This has been confirmed by results from different measurement campaigns. However, dual-slope exponential models would be more accurate to represent scenarios with large number of scatterers [29]. One shortcoming of these models is that they assume that the first path is always the strongest path which may not be true in general, especially for O2I and even I2I scenarios. In [30], the authors propose a model to accommodate the case when the first path may not be the strongest path. However, further independent measurement-based results would be necessary to verify the
validity and the choice of the model parameters.
电力延迟(PDP),之前最相关的文献中采用斜坡指数PDP工作,包括3 gpp SCM模型[28]。这已经被从不同的测量结果证实运动。然而,双斜率指数模型更准确代表场景与大量的散射[29]。这些模型的一个缺点是,他们假设第一个路径永远是最强的路径可能不是真的,特别是对于O2I甚至I2I场景。在[30]中,作者提出一个模型来适应情况下当第一个路径可能不是最强的路径。然而,进一步独立计量的结果将是必要的来验证模型的有效性和选择参数。
Doppler spread 多普勒传播
Dual mobility causes different Doppler characteristics compared to the traditional cellular links. Therefore Doppler spectrum for link level simulations [31] and 3GPP SCM [28] for system level simulations should be modified to reflect the changing of Doppler conditions. The D2D links experience larger Doppler spread which is caused by mobility of UEs at both sides and by the moving surrounding objects. In the literature, V2V channels are commonly characterised by means of RMS delay and Doppler spreads [32],and the Doppler power spectral density. Regarding the RMS delay spread, the smallest value is obtained in rural environments [33],and the largest value is obtained in urban environments [34].
双流动导致不同的多普勒特征相比,传统的蜂窝连接。因此多普勒频谱对链路级仿真[31]和3 gpp SCM系统级模拟[28]应该被修改,以反映变化的多普勒条件。D2D链接经历较大的多普勒扩散引起的流动性问题的双方和周围的移动对象。在文献中,V2V渠道通常为特征的RMS时延和多普勒扩散[32],和多普勒功率谱密度。关于RMS时延扩展,获得的最小值是在农村环境[33],在城市环境中获得最大的价值[34]。
Angle spread 角传播
In D2D communications, the link is symmetric, so both transmitter and receiver see a similar environment, and should have similar distributions for all parameters [35]. In typical cellular scenarios,the statistical angular properties of AoD (BS side) and AoA (UE side) are different [36]. Typically, AoD has smaller angle spread to account for the fact that the BS has higher antenna height and is further away from scatterers surrounding the user terminal [28, 37,38]. In D2D studies, UEs have similar antenna height and thus are likely to have similar scattering environments. To account for this fact, the D2D link in terms of AoD and AoA statistics is symmetrical and uses the parameters recommended for AoA, that is UE side, at both sides of the link [23].
D2D通信的链接是对称的,所以发射机和接收机都看到类似的环境中,并且应该有相似的分布参数[35]。在典型的蜂窝场景中,大气气溶胶的统计角度特性(BS)和AoA(问题)是不同的[36]。通常,大气气溶胶有小角蔓延至占BS具有更高的天线高度和远离周围散射用户终端(28日37、38)。在D2D研究中,问题也有类似的天线高度,因此可能会有类似的散射环境。考虑到这一事实,D2D链接在大气气溶胶和AoA统计信息是对称的,并使用参数推荐AoA,问题方面,双方的联系[23]。 3 D2D channel models
3 D2D信道模型
According to the analysis in Section 2, compared with conventional cellular channels, D2D channels have unique characteristics. This results in that modelling approaches and corresponding channel models for cellular systems cannot be directly used for D2D systems. Unlike a rich and fascinating history of the research in cellular channels, the investigation of D2D channel
modelling is still in its infancy.
根据第二部分的分析，与传统的蜂窝信道相比，D2D信道有其独特的特点。这导致建模方法和相应的信道模型不能直接用于D2D蜂窝系统。蜂窝信道有着丰富多彩的历史研究，与它不同的是，D2D信道造型仍处于起步阶段。
Based on the understanding of D2D propagation characteristics via either theoretical analyses and/or channel measurements, we can develop accurate yet easy-to-use channel models. We classify existing channel models in terms of their respective modelling approaches. The majority of existing channel models in the literature can be categorised as deterministic models and stochastic models, whereas the later ones can be further categorised as geometry-based stochastic model (GBSM) and measurement-based pseudo-geometric model (MBPGM) (Table 3).
基于对D2D传播特性的理解，通过理论分析和信道测量,我们还可以开发精确的易于使用的信道模型。根据各自的建模方法我们对现有的信道模型进行分类。著作中现有的大多数信道模型可以归类为确定性模型和随机模型，而后来的可以进一步归类为几何随机模型(GBSM)和伪几何计量模型(MBPGM)(表3)。
Deterministic channel models characterise D2D channelparameters in a purely deterministic manner, for example, ray-tracing approach [39]. This modelling approach needs an accurate database, high computation time and use approximations of the Maxwells equation for its solution [40]. Therefore deterministic channel model typically has high complexity and cannot be easily generalised.
确定的信道模型以确定性的方式描述D2D信道参数，例如射线跟踪方法。这种造型方法需要一个准确的数据库、大量的计算时间以及使用麦克斯韦方程的近似解决方案。因此确定的信道模型通常具有较高的复杂性并且不易被普及。
The MBPGM is entirely based on channel measurements. Examples of MBPGM include the widely used SCM and WINNER models, as well as the recently developed COST 2100 model [41]. More recently, by considering D2D communication environments, 3GPP proposes D2D MBPGMs [42] based on the modifications of WINNER model [24]. However, owing to the nature of MBPGM, the 3GPP D2D channels model cannot easily mimic some important characteristics owing to the dual mobility, for example, scatter drifts and transitions between different propagation environments or between LoS and NLoS scenarios. Note that the widely used SCM, COST 2100 and WINNER models are often mistakenly referred to as GBSMs.Such misunderstanding is largely due to the fact that the main channel parameters are all related to scatterers/clusters, for example, AoA/ AoD and angel spread. However, note that all these scatterer/ cluster-related parameters are actually obtained solely based on measurements, instead of predefined stochastic distributions of the scatterers/clusters. Therefore all the SCM, COST 2100 and WINNER models are more properly classified as stochastic parameter models. In the following, we will give a brief introduction of those widely used MBPGMs.
伪几何计量模型完全基于信道测量。MBPGM的例子包括广泛使用SCM和WINNER模型以及最近发展的COST2100模型。最近,通过考虑D2D通讯环境,3 gpp基于修正后的WINNER模型提出D2D伪几何计量模型。然而,由于MBPGM双重流动特性,3 gpp D2D信道模型无法轻易模仿一些重要特征，例如,不同的传播环境之间以及仿真结果之间的散射漂移和转换。值得注意的是，广泛使用的SCM,COST2100和WINNER模型经常被错误地称为几何随机模型。这种误解很大程度上源于这样的事实,主要信道参数都与例如AoA /AoD以及角差这样的散射/集