Multi-Code TDMA (MC-TDMA) for Multimedia Satellite Communications
用于多媒体卫星通信的MC--TDMA（多码时分多址复用）
R. Di Girolamo and T. Le-Ngoc
Department ofa Electricl and Computer Engineering - Concordia University
1455 de Maisonneuve Blvd. West, Montreal, Quebec, Canada, H3G 1M8
ABSTRACT
摘要
In this paper, we propose a multiple access scheme based
on a hybrid combination of TDMA and CDMA,在这篇文章中，我们提出一种基于把时分多址复用和码分多址复用集合的多址接入方案。 referred to
as multi-code TDMA (MC-TDMA). 称作多码—时分多址复用The underlying TDMA
frame structure allows for the transmission of variable bit
rate (VBR) information,以TDMA技术为基础的帧结构允许传输可变比特率的信息 while the CDMA provides inherent
statistical multiplexing.和CDMA提供固有的统计特性多路复用技术 The system is studied for a multimedia
satellite environment with long-range dependent
data traffic,and VBR real-time voice and video traffic研究这个系统是为了在远程环境下依赖数据传输和可变比特率的语音和视频传输的多媒体卫星通信系统 . Simulation
results show that with MC-TDMA, the data packet
delay and the probability of real-time packet loss can be
maintained low. 仿真结果表明：采用MC-TDMA的多媒体卫星通信，数据包延时和实时数据丢失的可能性可以保持很低。The above advantages are achieved at the
expense of soft blocking, which occurs when many packets
are transmitted simultaneously, but on different spreading
codes.上述的优点是以软阻塞为代价达到的，当很多数据包以不同的扩频码同时传输时会发生软阻塞。 Similar conclusions are drawn for simulations
where the channel propagation conditions, adjacent beam
interference, and imperfect power control are considered.当仿真时考虑到信道传播条件，相邻的电波干扰和非理想功率控制等因素也会有相似的结论。
1 INTRODUCTION介绍
The nature of satellite channels make them ideally suited
for broadcast and multipoint transmissions. 卫星信道的性质非常适合广播和多点传输The envisioned
satellite systems do not only provide a backup to
the terrestrial wire and wireless networks, but in many situations
they are the only feasible alternative.未来卫星系统不只是给地面上有线和无线网络提供一个备份，在很多状况他们还是唯一可行的替代 This is particularly
true in cases where the infrastructure基础设施 to support the
terrestrial systems is not in place. 在支持地面的系统的基础设施不到位的情况下上面的说法就更正确了In such environments,在这种情况下
the satellite provides a means of communication which is
insensitive to the location and the distance between the
communicating users.卫星会提供一种通信方式，这种通信方式对在通信的用户所在的位置和他们之间的距离不敏感
With the increased emphasis placed on multimedia, 现在重点放在多媒体上，any
proposed systemmust be able to support a mixture of voice,
video, and data. 任何被提议的系统都必须可以支持语音，视频和数据的混合Typical applications could include video
conferencing, LAN interconnection, medical imaging, remote
video monitoring, and the like典型的技术包括视频会议，局域网互联，医学影像，远程视频监控和类似于这类的技术。. As a consequence, the
satellite system must be able to support all these services，and any future services which may become popular.所以，卫星系统必须可以支持所有这些服务和未来可能很受欢迎的服务。The problem lies in the characterization of this multimedia traffic
and in the different quality of service (QoS) requirements
for each of these. 问题就在于这种多媒体交换的描述和服务质量的不同要求不同。Packetized data, for instance, is generally
jitter tolerant but loss sensitive, 比如，封包化的数据通常可以容错但是不够敏感。and often requires the retransmission
of packets which are received incorrectly.而且经常要求重新发送在接收端被错误接受的数据包 In addition,此外 recent traffic studies have shown that data traffic
exhibits long-range dependence [ll. 最近在通信交换的研究表明数据的交换存在一种长期的依赖，On the other hand,另一方面
voice and video are delay sensitive (real-time), 语音和视频信号对延迟是敏感的，but normally
can sustain a certain level of packet loss.但是一般情况下他可以支撑到数据包丢失的某个电平 Furthermore,此外 as
a result of the coding techniques used,作为编码技术被利用的结果 the real-time traffic
can be of variable bit rate (VBR). 它可以实现可变比特率的实时交换Any designed system
should address the above issues.任何一个设计的系统都应该考虑上述问题。
The second issue to be resolved is the method of multiple
Access.第二个要解决的问题就是多址接入的方法 How to geographically dispersed earth stations(or users) supporting multimedia traffic, 在地理上怎么划分可以支持多媒体交换的区域或者用户efficiently share the available fixed uplink bandwidth怎么有效的分享可利用的固定的上行带宽. One of the
simplest approaches is to use a form of demand assigned
TDMA or multi-frequency TDMA (MF-TDMA),where an
earth station periodically requests capacity, 最简单的方法就是使用一种按需求分配的形式的TDMA或者多频 TDMA。地面的无线站周期性的请求信道带宽 ，有一个中心程序机来响应这些请求and a central
scheduler honors these requests [2]. A second broad class
of technique which has recently been gaining momentum,
has CDMA as a basis. 最近已经蓄势待发的以CDMA为基础的第二大类技术，These spread spectrum approaches
require very high transmission chip rates, and are hampered
with problems dealing with synchronization and unequal
power control.这些扩频方法要求很高的传输芯片率，也受到处理同步和不平均的功率控制 等问题的限制However,然而它们拥有很多优点包括固有的统计复用， they possess a number of advantages,
including inherent statistical multiplexing, a graceful
degradation in system performance as load increases,当负荷增加时在系统性能上有一个故障弱化的优点
coexistence with narrowband analog systems in the same
frequency band, etc.在相同的频带内可以和窄带模拟系统共存。 [31. There are numerous variants of
CDMA which are capable of dealing withVBR traffic. These
include multi-code CDMA 141, and variable spreading gain
CDMA[5]. 为了处理可变比特率的传输问题，存在许多码分多址接入方法的变形，包括多码的CDMA和可变的扩频增益CDMA 。
In this paper, we propose and investigate the performance
of a combined TDMA/CDMA technique, which we
refer to as multi-code TDMA (MC-TDMA).以上内容我们提出和研究了把TDMA和CDMA结合的技术的性能，就是我们称之为的MC-TDMA。The figures of
merit used to study performance include packet delay of
the long-range dependent data traffic, packet loss probability
of the VBR voice and video, and the probability of soft
blocking, which is defined in the next section.在接下来的章节我们讨论的是该系统的优点包括依赖数据的远程传输的数据包延时，丢失可变比特率的语音和视频数据包的可能性，和发生软阻塞的可能性，通过这些来用来研究该系统的性能
2 PROPOSED MC-TDMA SATELLITE SYSTEM
A typical configuration of the proposed system is shown
in Figure 1. 图1是所提出系统的典型结构。The satellite has multibeam capabilities, with
earth stations in different spot beams being able to communicate
via the on board switch. 卫星有多电波功能，可以通过交换机使地面的基站在不同地方的电波进行通信。Since we are only interested
in the performance of the uplink access, in this paper we do
not consider any switching. This allows us to focus only on
the earth stations of a single spot beam.开始我们只对上行接入性能感兴趣，这章我们不考虑交换，所以我们可以专注于单点波束的地面接收站。
Within this selected spot beam there are LU earth stations,
each of which has a pre-assigned spreading code,
denoted by the set {C;} i = 1,2, ..., Lu. These codes act as
user IDS,这个被选定的单一波束在地面有LU个地面接收站，每个基站都预分配了扩频码，被记为字母{C；} i=1,2,…..LU，这些码作为用户的ID。since no two users in the same spot beam have
the same spreading code. 所以两个用户不会有相同的扩频码。Each earth station transmits
its packets by spreading the information over the available
bandwidth, using its own user ID code. 每个基站用每个用户拥有的ID码使用空闲的带宽传输数据包At the satellite, the
combined signal from all active users goes through a multiuser
receiver. 在卫星上，使用者的合信号经过一个多用户接收机The simplest form of such a receiver is a bank of despreading correlators形式最简单的接收机就是一个有解扩因子的存储器。
. The satellite then switches each
received packet to the appropriate downlink beam,然后卫星把每个接收到的数据包传给适当的下行波束。 where
these are time division multiplexed (TDM) onto a carrier.下行波束在载波上使用的是TDMA
Since the hardware on board the satellite is expensive, 因为卫星上的硬设备比较昂贵，所以要限制多用户接收机的大小the
multi-user receiver will have to be limited in size. Here, we assume that for each spot beam, the receiver can handle at
most LR active users, where LR << Lu.现在我们假设对于每个波束，接收机最多可以处理LR个用户，LR要小于LU
2.1 Uplink Access and Scheduling上行接入与调节
One of the unique features of MC-TDMA deals with the
method of scheduling.用调节的方法解决问题是MC-TDMA的一个特色 The simplest way to highlight this uniqueness, is by means of comparison to a demand assigned
MF-TDMA system.强调这个特性最简单的方式就是用他和按需求分配的MF-TDMA系统做对比。 The first step in initiating a call
is to send a request to the scheduler by means of some reservation
channel. 初始化一个通话的第一步就是通过一些保留的信道发送一个请求给调度者Those earth stations having calls which
have been accepted are classified as active users.已经被基站接受的通话过程被定义为使用者 Obviously
for MC-TDMA, the number of active users per spot beam
is confined, by hardware, to be less than LR,显然对于MC-TDMA系统来说每束光波的使用者被硬件限制在LR个之内， the maximum
number of despreading codes allowed per spot beam.LR是每束光波所能允许的最大解扩码的数量。
Figure 2 shows a side-by-side comparison of the division
of the uplink bandwidth for MF-TDMA and MC-TDMA.图2是MF-TDMA与MC-TDMA的上行带宽的对比
The hybrid TDMA part of both techniques imposes an underlying
frame structure. 这两中技术混合TDMA的部分都运用了基本的帧结构Each of these frames is subdivided
into a constant number of recurring slots, which can
accommodate a fixed size packet (or cell). 每个帧被细分到一个有连续序号的循环槽中，这个循环槽可以容纳固定大小的数据包We assume that
all information to be transmitted is first packetized into
ATM type cells.我们假设所有的信息都以异步传输模式（ATM）打包成数据包再被传送。 A user varies his transmission bit rate by
transmitting more or less packets within a frame. 用户可以通过改变一个帧中数据包的数量来实现可变的比特率传输。However,
the cell rate (in packets per frame) for each earth station
will be limited to some maximum (Rmax,)然而，每个基站数据包的速率将会被限制在一个最大值Rmax之内. If an earth station's
instantaneous cell rate exceeds this maximum, then
one of two things happens:如果基站的瞬时速率超过这个最大值时，以下两件事会发生一件。一，当实时数据包不能承受任何时延抖动，那么他们就不能进行排队发送，所以数据包是实时的，传输容量又达到了最大值，那么超过的数据包会丢失。二，如果数据包混合了实时数据和抖动容忍数据，那么实时数据会优先发送，剩下的时隙用了传输抖动容忍数据包，那些不能被传输的数据包排队等待，他们只能竞争随后帧的容量，这种方式的数据传输类似于ATM服务中的可用比特率传输。
1. Since real-time packets cannot sustain any delay jitter,
they cannot be queued. As a result, if the packets are realtime,
then up to the maximum are transmitted (Rmax), and
any excess packets are rejected (lost).
2. If the packets are a mixture of real-time and jitter-tolerant,
then priority is given to the real-time packets. Any remaining
slots are used for the jitter-tolerant data traffic. Those
data packets which cannot be transmitted are queued, and
contend for capacity in subsequent frames. The treatment
given to the data traffic resembles the available bit rate
(ABR) ATM service.
In both cases, the number of slots in the TDMA frame is
equal to the maximum cell rate of an earth station
For MF-TDMA, each earth station requests capacity from
the scheduler based on the number of slots required. 在上述两种情况下，TDMA帧结构中时隙的数量等于MF-TDMA基站的最大元组速率，每个基站根据时隙请求的数量向调度者请求容量。Based on the requests from all users, the scheduler then explicitly
informs each and every earth station the frequency-time
channels which have been assigned to it.根据所有用户的需求，调节器明确的指出每个基站已经被分配的跳频扩时信道 In assigning these
frequency-time channels, the scheduler gives priority to the
requests from the real-time packets, and it ensures that no
earth stationis assigned more than R,,, channels. 在分配这些跳频扩时信道时，调度者会优先响应实时数据包，而且要保证每个基站分配的信道都小于Rmax。This latter
condition guarantees that no earth station will transmit
on two different carriers at the same time. 后者的情形保证了每个基站都不能同时传输两个不同的载波。Notice that an
earth station must request capacity, and the scheduler must
assign capacity, on a frame-by-frame basis.注意到每个基站都必须请求容量，而且调度者必须一帧一帧的给他分配容量， In contrast,
for MC-TDMA the scheduling process is greatly simplified.相反的，对于MC-TDMA系统这个调度程序就大大的简化了。
Once a call is accepted, the scheduler notifies the earth station
which requested the connection, say station j , that it
may begin transmitting its information by using its user ID
spreading code Cj. 一旦一个通话被接收，调度者通知被要求连接的基站，叫做基站j,他就开始通过使用用户的ID解扩码C进行传输信息。Secondly, it dynamically sets one of the
despreading codes on board the satellite to Cj, in preparation
for the transmissions from earth station j.然后他动态的把解扩码传输给卫星，为基站j传输信息做准备。Once this is done, the scheduler involvement is complete.做完这步，调度者的工作就完成了。 Each earth
station then randomly transmits its packets in one of the
available TDMA slots.然后每个基站通过可用的TDMA时隙随机的发送数据包 In summary, we can say that the
scheduler operates on a frame basis for demand assigned
MF-TDMA, but on a call basis for MC-TDMA. 总结下，我们可以说对于以需求分配为基础的MF-TDMA系统调度者是作用在帧结构上，但是对于MC-TDMA调度者是作用在呼叫上的Besides the obvious savings in on board hardware owing to the reduced
scheduler complexity, 此外，因为减少调度者的复杂度明显的减少了再硬件上的开支。there is also the added advantage of
a shorter call set-up time.同样对于建立一个更短时间的呼叫时间上有更大的好处。
A second major difference between MF-TDMA and MCTDMA
concerns the orthogonality of user transmissions.MF-TDMA和MC-TDMA之间还有第二个主要的区别就是用户传输时的正交性。
With MF-TDMA, each frequency-time channel is orthogonal,
and as a result at the satellite receiver, the transmitted
packets can be treated independently. 在MF-TDMA中，每个跳频扩时信道都是正交的，所以在卫星接收端，传输的数据包可以独立的处理。This is not the case
in MC-TDMA, where the packets transmitted in the same
slot (each on their own spreading code) act as interference
to one another (multiple access interference - MAI). 对于使用相同时隙传输数据包的MC-TDMA就不能这样做，因为同一个时隙的信号对另一个信号是一种干扰（多径干扰）The number of packets that can be transmitted simultaneously
is limited by the allowable MAI.在可允许的多径干扰范围内可以同时传输数据包 Given that the processing
gain for each earth station is PG, and that the number of
users in a slot is N, an approximation to the multiple access
interference power spectral density is found in [61假设每个基站为Pg,一个时隙的用户数为N，多径干扰的频谱密度的近似计算公式为
(l)th e MA1 increases as N increases, which in effect reduces
the received signal-to-total noise ratio (SNR,,,).MAI随着N的增加而增加。可以较少接收端的信噪比Normally, a threshold received signal-to-multiple access interference
ratio (Sh9AIRthTeshe)x ists, for which the signal quality
cannot be guaranteed if the received SMAIR falls below
this threshold. This occurs when N is greater than some
N,,,. These packets are not lost, but rather, we say that
they are soft blocked [7]. Notice that when soft blocking
does occur, all packets within a slot are affected.一般的，在接收端会有一个信干比的门限值，如果接受到信干比减到低于这个门限值，那么信号的质量不能保证。当N大于Nmax这种状况会发生。这些数据包不会丢失，更确切的说这就是软阻塞。值得注意的是当发生软阻塞时，在同一个时隙的所有数据包都会受到影响。
3 SIMULATION MODEL AND RESULTS仿真模型和仿真结果
The simulation results were obtained for various combinations
of input parameters, using the OPNET simulation
tool. 使用OPNET仿真软件，通过输入不同组合的参数可以获得仿真结果。The earth stations were assumed to be homogeneous, and the load of each was fixed.地面上的基站假设是同种类的，而且他们的负载是固定的。
The total system load was
varied by changing the number of active users. 整个系统通过改变使用者的数量来改变负载。As already
mentioned, we did not model the call admission problem.上面已经提到，我们不模拟呼叫接入的问题
Consequently, all simulation runs were performed assuming
a fixed number of active users.所以，所有的仿真都假设是以固定的使用者进行的。
The downlink TDM payload per spot beam was taken
to be 4096 packets/frame, and we assumed a bandwidth
of 95 MHz.它的下行链路的时分多路复用每波束的净负荷为4096个数据包或者帧结构。而且我们假设有95M的带宽。 The maximum load on the uplink was also
taken to be 4096 packets per frame, and we normalized all
load values to this maximum.上行链路的最大负载也认为是每帧4096个数据包，我们把所有的负载都标准化到这个最大值。Furthermore, we assumed
an average cell rate per earth station of 48 packets/frame,
and a peak cell rate of Rmax = 128 packets/frame. 此外，我们假设每个基站的元组速率是48个数据包或者帧结构。而且最大的元组速率为Rmax=128个数据包或元组In
all, four load points were considered - 0.95,0.8,0.5, and 0.2.总的说来，有四个负荷点分别是- 0.95,0.8,0.5,和0.2.
These corresponded to 81,68,43, and 17active earth stations
respectively.这个数据分别符合81,68,43，和17个使用者的基站 For each load point, two different traffic mixes
were examined - voice dominant and data dominant. 对于每个负载点，语音检测点和数据点是混合两个不同的传输方式For the data dominant case, the ratio of voice/video/data traffic
for each earth station was 20%/10%/70%, while the ratio
for the voice dominant case was 70%/10%/20%.对于数据主导的情况，每个基站的语音，视频和数据流量所占的比例是20%/10%/70%，当语音主导的情况下，语音，视频和数据流量所占的比例是70%/10%/20%。
3.1 Traffic Models
As can be seen from Figure 1, 从图一可以看出，we assumed each earth
station had a single voice, video and data source. 我们假设每个基站都是单一的语音源，视频源和数据源。The voice
and video sources were modelled as independent two state
Markov Modulated Poisson Processes (MMPP's).语音信息是视频信息按两个独立状态的 马尔可夫调制泊松过程模拟。 Results
in the literature, show that MMPP sources are quite effective
in modelling single video and multiple voice sources .文献中的结果表明经过马尔可夫调制泊松过程的信息源在模拟单个视频信号或者多个语音信息源非常的高效。
MMPP sources are characterized by exponential sojourn
times, and Poisson arrivals in each state. 经过马尔可夫调制泊松过程的信息源以指数次为特征，泊松到达每一个状态。The parameters
for each of the two state processes were selected to yield a
high level of burstiness, and to match the required load.为这两个状态过程的选择的参数有高的突发性。而且要匹配要求的负载The data, on the other hand, was generated by a Pareto Modulated Poisson Process (PMPP) matched to yield a Hurst
parameter of 0.8 另一方面，通过帕累托调制泊松过程产生的数据要匹配产生一个赫斯特0.8的参数。The PMPP is also a two state process
with Poisson arrivals in each state. 帕累托调制泊松过程也是两种状态的过程 ，泊松过程可以到达每个状态。However, unlike the
MMPP, the sojourn times are Pareto distributed. As shown
in (91, the PMPP captures the long-range correlations which
were reported in (11.然而，不像MMPP，寄居的时间服从帕累托分布，在参考【9】中表明。PMPP可以获得参考【11】中提到的远程相关性。
As is shown in Figure 3, the use of the above mentioned
correlated traffic models has a profound effect on the performance
of the proposed MC-TDMA system, especially
at high loads.如图3所示，以上提到的相关传输模型对MC-TDMA的性能有深远的影响。特别是在高负载的情况下In this figure, we show the Pr[soft blocking]
versus the processing gain, for both correlated (data
dominant) traffic and independent traffic.在图中，不管是相关的传输还是独立的传输，表明了软阻塞对抗增益处理。 We assumed a
SMAIRthTesho f 6 dB, and determined the probability that
a packet would be soft blocked. For the correlated traffic,
we resorted to simulation. For the independent traffic, on
the other hand, the results were obtained analytically. This
was possible, since the number of active earth stations in a
slot has a binomial distribution.我们假设SMAIR为6dB，我们设定一个数据包为软阻塞的概率，对于相关的传输，我们求助仿真，另一方面，对于独立的传输，结果可以分析得到。这是可以实现的，如果基站的一个时隙中使用者的数量服从二项分布。
Notice that at 20% load, the Pr[soft blocking] is almost
the same for both the correlated and the independent traffic
cases.注意到如果是20%的负载，软阻塞对于相关传输和独立的传输的情形是一样的 However, the curves begin to diverge as the load
increases. 然而，当负载增加时曲线开始偏离。As an example, for the 80% load case and at a
processing gain of 115, the independent traffic model underestimates
the Pr[soft blocking] by about 2 orders of magnitude比如，当时80%的负载的情形下，且处理增益为115时，独立的传输模型下软阻塞会低估了大约两个数量级。
(6 x as compared to 6 x low7). The above clearly
shows the need for using a correlated traffic model when
dealing with multimedia traffic, and consequently, all subsequent
results assume this type of traffic.以上清楚的表明当处理多媒体传输时我们需要使用相关的传输模型。所以，随后的结果都假设是这种类型的传输。
3.2 Simulation Results Obtained得到的仿真结果
The figures of merit chosen to study the performance of
the MC-TDMA system include data packet delay when a
data packet cannot be transmitted owing to lack of uplink
capacity and is queued; the real-time loss probability which
occurs when a real-time packet is lost; and the soft blocking
probability.选择表中的优点来研究MC-TDMA系统的性能包括当一个数据包因为缺少上行容量不能被传输时的排队数据包延时，实时数据包丢失发生的概率和发生软阻塞的概率。
Unlike the demand assigned MF-TDMA system, where
a centralized scheduler controls the transmission for each
earth station, MC-TDMA can be thought of as providing a
distributed control. Each earth station transmits its packets
independently of other earth stations. As the total load increases,
the number of active users increase, but the load per user remains the same.不像按需求分配且有一个中心调度者来对每个基站的传输进行控制的MF-TDMA系统，MC-TDMA系统提供一个分散的控制，每个基站独立的传输他的数据包，当总的负载增加时，使用者的数量也增加，但是每个用户的负载保持不变。Since each of these users is transmitting
independently, the increases in system load should not
effect the data packet delay performance. 当每个用户都独立的传输，系统负载的增加不会影响数据包时延的性能。On the other hand
for MF-TDMA, as the total system load increases, there are
less frequency-time channels to be assigned, and more and
more data packets will have to be queued. 另一方面，对于MF-TDMA系统，当整个系统负载增加时，那么更少的跳频扩时信道被分配，越来越多的数据包会排队等待发送。The result is
an increase in the data delay as the load increases. These
features are observed in Figure 4, 所以负载增加数据时延也会增加。这些特性在图4中很清楚的体现了。where the earth station
queueing delay is plotted versus the total system load. 该图绘出了基站的排队等候延时与整个系统负载的对比。
For MC-TDMA, the delay is equal to half a frame time for all
loads, and for both voice and data dominant traffic cases.
As a comparison, we have also included the simulation
results for a combined free/demand assigned MF-TDMA
system . 对于MC-TDMA系统，所有负载的时延是帧传输时延的一半，对于语音和数据传输都是这样的。作为对比，我们也仿真了结合自由分配和按需求分配的MF-TDMA系统的结果。As expected, for the MF-TDMA case, the data
packet delay increases with load, and with percentage of
data traffic.正如我们期望的那样，MF-TDMA系统，数据包的延时随着负载的增加和数据流量比例的而增加。
In terms of real-time packet loss, for both the MC-TDMA
and MF-TDMA systems simulated, no packets were lost for
the some lo6 packets generated.对于实时数据包的丢失，当产生个数据包时对于MC-TDMA和MF-TDMA系统仿真结果都是不会丢包。
The benefits noted above are achieved at the expense of
an increase in soft blocking. 上述说的好处都是以增加软阻塞为代价而实现的。Since MF-TDMA transmitted
packets are completely orthogonal, no equivalent to soft
blocking occurs in this system. Unfortunately, soft blocking
does pose a problem to MC-TDMA type systems. 虽然MF-TDMA传输数据包时是完全正交的，这个系统不会发生软阻塞。不幸的是软阻塞会对MC-TDMA产生影响。Figure
5 shows the Prrsoft blocking] versus the processing gain,
for bath the voice and data dominant traffic cases. The
results show that at high system loads, the traffic mix has
a pronounced effect on soft blocking. For the traffic mixes
we considered, the probability was much higher for the
burstier data dominant case. This result should have been
expected based on the conclusions from Figure 3.图5显示的是对语音和数据传输时软阻塞对抗增益处理。这个结果表明在高系统负载的情况下，传输的混合对软阻塞有明显的影响。我们认为对于混合传输，发生软阻塞的概率比突发数据传输的情况要大。这是基于图3的结果做出的结论。
Since soft blocking can be very high, any approach to reduce
it would be beneficial. 当软阻塞发生概率很高时，任何可以减少它的方法都是有用的。Some ideas on how this can be
achieved become apparent when we plot Pr[soft blocking]
versus the SMAIRthreah, as shown in Figure 6. Notice that
as the threshold decreases, so does the soft blocking. 如图6中表明为什么有些想法可以明显的达到这个效果。Consequently,
the following comments can be made about soft
blocking.所以，下面可以做一些关于软阻塞的结论。
1. When we plotted the curves for Figures 3 and 5, we
treated all packets equally. However, voice, video, and
data each have their own QoS requirements, and as a result,
each has aRthresh. Voice, for instance, can be
received correctly at levels of MA1 which would be unacceptable
for data, since the latter is loss sensitive.当我们在绘制图3和图5时，我们假设所有的数据包都是相等的，而且语音，视频和数据都有他们自己的QOS要求。所以每个都有SMAI，比如，在MA1电平上语音信号可以准确的接受，但对于数据就不能，因为数据更容易丢失。
2. Although we have not mentioned forward error correction
FEC coding, CDMA systems can be heavily coded, with no increase in bandwidth. The added protection provided by the coding could then be used to reduce
虽然我们没有提及向前纠错，FEC码，CDMA系统可以实现不增加带宽的重编码，通过编码增加了保护作用从而可以减少软阻塞。
3. All soft blocking performance results were obtained assuming
a simple receiver (a bank of despreading correlators).
In fact, there are a number of more efficient multi-user
receivers which will yield the same performance at lower
SMAI所有的软阻塞性能结果都是假设一个接收端得来的，实际上，有很多更高效的可以以低的SMAI产生相同性能的多用户接收端。
4. Alteratively, when soft blocking is high, we could use
some form of flow control, whereby we queue the jitter-tolerant
data packets at the earth stations. This would reduce
the number of packets transmitting simultaneously.当软阻塞发生概率比较大时，我们可以用一些流量监控的办法，我们可以让容忍抖动的数据包在基站排队等待，这样可以减少同时发送数据包的数量。
3.3 System Imperfections系统缺陷
Since CDMA systems are interference limited, it is essential
to consider system imperfections when talking about
performance.自从CDMA系统有干扰限制后，在谈论系统性能的同时有必要考虑系统的缺陷。 Briefly, in this paper we focus on three major
imperfections - namely signal propagation impairments,
adjacent spot beam interference, and imperfect power control.简单来说，这章我们集中注意在三个主要缺陷上，分别是：信号发送障碍，临近波束干扰和不完美的功率控制。
To study these impairments, we fixed the signal-to thermal
noise ratio Eb/No to 12 dB, and observed the received
SNR for each transmitted packet. 为了研究这三个缺陷，我们定义信号到终端的噪声比例为Eb/No 等于12dB。和观察接收端对每个传输数据包的信噪比。For each slot, the
received signal-to-totalnoise ratio for earth station i is given
对于每个时隙，我们给出每个基站的每个接收端的信号到终端的噪声比例为：
where E& is the received signal energy from user i, No is
the thermal noise PSD, and &,MA1 and &AB1 are the multiple
access interference PSD and adjacent beam interference
PSD, respectively. E是从用户I那接受到的信号能量，NO是终端的噪声，&,MA1 和 &AB1是各自多址接入干扰PSD和临近的波束干扰PSD。 These latter terms are given by:后面这些数据有一下公式求得：
In the above, L denotes the number of earth stations transmitting
simultaneously, E&,i) denotes the received signal
energy from user k in adjacent beam j, K denotes the number
of adjacent beams with the same uplink frequency, Ljdenotes the number of earth stations transmitting simultaneously
in adjacent spot beam j , and ,6' denotes the attenuation
provided by saiellite antenna to adjacent beam earth
stations.L是同时在传输数据的基站数，是接受到的临近波束i的用户k的信号能量，K是有相同上行频率的临近波束的数量，
是临近波束j同时传输数据的基站数，
是卫星天线对相邻波束基站产生的衰落。
Propagation Channel [lll: We assumed a quasistatioAa&
channel model, governed by a 7-state Markov
chain. where each state denoted a sDecific weather condition
- clear sky, rain, cumulus cloud, thunder shower, overcast,
intermittent light rain. The duration of a state was
assumed to be exponentially distributed with an average of
6 hours.传播信道：我们假设一个半稳定的信号模型，它被有七种状态的马尔科夫链控制，每个状态用一个特定的气象信息标记，比如晴朗，下雨，多云，雷阵雨，阴天，间歇性小雨。每个状态的持续时间我们假设为平均6小时。
Adjacent Beam Interference: When the uplink frequency
band is reused in adjacent spot beams, earth stations in
these spot beams act as interference to the transmissions
in the desired spot beam. The satellite antenna pattern
provides some discrimination against these adjacent beam
earth stations. We assumed an average case, where all
earth stations in the adjacent beams were located in the
center of that spot beam. We further assumed that the
satellite antenna provided 17.5 dB of discrimination to the
transmissions from these earth stations. For the purposes of
simulation, we considered three frequency reuse scenarios相邻波束干扰：当上行频带被相邻波速重新利用，在这一波束内的基站对于想要被传输的那个波束是一种干扰，卫星天线模式提供了一些可以识别临近波束干扰的基站的方法。我们假设一种平均状况，所有临近波束的基站都设立在这个波束的中心，再假设我们的卫星天线对这个基站提供17.5dB辨识度，这种情况下的仿真，我们认为三个频段可在不同的地方重新使用。见图7（See Figure 7).
Power Control: The purpose of power control is to equalize
the received power of each of the transmitting users.
In our case, the received power from each of the users is
different since the signals are randomly attenuated by the
channel. To deal with power control errors, we used a
behavioural model, in which the power control error was
approximated by a log-normally distributed random variable.
This approach is used quite often in the analysis of
cellular systems , and we extend the results to the satellite
system.功率控制：使用功率控制的目的是补偿接收端从传输用户那接受的信号功率。就我们而言，因为信号在信道中会随机的衰减，所以每个用户的接受功率是不同的，为了处理功率控制误差，我们使用自适应模型。这样功率控制误差大约服从对数正态分布，这种方法在分析蜂窝系统时经常用到。我们在卫星系统中沿用了这个结果。
The results of the simulations for a 95% load data dominant
case, are shown in Figure 8. In this figure, the frequency
of occurrence of SNR,,, is plotted, for the three
different frequency reuse scenarios. Given a minimum required
SNR threshold, these curves can be used to determine
the Pr[soft blocking]. From the graph, it seems clear
that the received SNR is maximized by using the largest
possible processing gain, even if this means more adjacent
beam interference.如图8所示，它显示了95%负载数据形式下的仿真结果，图中绘制了发生SNR的频率图，对于这三个不同频率使用地点，给出了一个最小的SNR门限值，这个曲线可以判决软阻塞，从图可以看出，接收的信噪比会变大当使用最大可能的增益，尽管这将意味着有更多的临近波束的干扰。
4 CONCLUSIONS结论
In this paper, we have proposed and described a MCTDMA
satellite svstem for multimedia traffic，We
have tried to highlight some of the features of MC-TDMA,
such as the reduced scheduler complexity, and the need
for a large transmission bandwidth to maximize processing
gain. 0% the performance side, we simulated the satellite
system, and observed that the data delay can be maintained
to a minimum of half a frame time. Furthermore, real-time
packet losses are very small for the proposed technique -no losses occurred during the simulation runs. This performance
is achieved at the expense of soft blocking, which
increases as the load increases.这篇文章，我们提出和描述了一个MC-TDMA用于多媒体传输的卫星系统，我们试着突出MC-TDMA的一些特性，比如说减少调度器的复杂度，为了增大增益处理需要更大的传输带宽，在性能方面，我们仿真了卫星系统，我们观察到数据时延可以维持在最小等于帧传输时间的一半。而且，对于我们提出的技术实时数据包的丢失率非常小，在仿真时没有发生丢包。这些性能是以会因为负载增大而增大的软阻塞为代价得以实现的
When system imperfections were considered, similar
conclusions were observed. In addition, the frequency
of occurrence of the received signal-to-noise ratio was obtained
for various frequency reuse scenarios. These curves
can be used to evaluate the Pr[soft blocking] if a threshold
SNR is known. The results also suggest that maximizing
the processing gain is more important than minimizing the
adjacent beam interference.当考虑到系统缺陷时，可以得到同样的结论。此外，通过不同地点再使用相同频率得到了接受端信噪比，这些曲线可以估算软阻塞，如果知道NSR的门限值，结果表明增大处理增益比减小临近波束干扰更重要。
References
[l] W.E. Leland, M.S. Taqqu, W. Willinger, D.V. Wilson, 'On
the Self-similar Nature of Ethernet Traffic (Extended Version),'
IEEEIACM Transactions on Networking, Vol. 2, No. 1,
pp. 1-15, February 1994.
[2] B. Jabbari and D. McDysan, ,Performance of Demand
Assignment TDMA and Multicarrier TDMA Satellite Networks,'
IEEE Jouml on Selected Areas in Communications,
Vol. 10, No. 2, pp. 478-486, February 1992.
[3] W.C.Y. Lee, 'Overview of Cellular CDMA,' IEEE Transactiuns
on Vehicular Technology, Vol. 40, No. 2, pp. 291-302,
May 1991.
[4] C.L. I, G.P. Pollini, L. Ozarow, R.D. Gitlin, 'Performance
of Multi-Code CDMA Wireless Personal Communication
Networks,' Proceedings VTC'95, pp. 907-91 1.
[5] C.L. I and KX. Sabnani, 'Variable Spreading Gain
CDMA with Adaptive Control for True Packet Switching
Wireless Networks,' Proceedings ICC'95, pp. 725-730.
[6] M.B. Pursley, 'Performance Evaluation for Phase-Coded
Spread-Sped" Multiple- Access Communication - Part 1 :
System Analysis,' IEEE Transactions on Communications, Vol.
[7] A.M. Viterbi and A.J. Viterbi, 'Erlang Capacity of aPower
Controlled CDMA System,' IEEE Joumal on Selected Areas in
Communicatim, Vol. 11, No. 6, pp. 892-900, August 1993.
[8] M. Schwartz, Broadband Integrated Networks, New Jersey:
25, NO. 8, pp. 795-799, August 1977.
Prentice Hall, 1996.
[9] S.N. Subramanian and T. Le-Ngoc, 'Traffic Modeling in a
Multi-media Environment,' Proceedings CCECE'95, pp.-838-
841.
[lo] T. Le-Ngoc and S.V. Krishnamurthy, 'Performance
of Combined Free/Demand Assignment Multiple-Access
Schemes in Satellite Communications', International Joumal
@Satellite Communications, Vol. 14, pp. 11-21/1996.
[ 111 C. Loo, 'Statistical Models for Land Mobile and Fixed
Satellite Communications at Ka Band,' Proceedings VTC'96.
本文来源：https://www.2haoxitong.net/k/doc/435c1586d4d8d15abe234e60.html
文档为doc格式