Professional Documents
Culture Documents
Planning of Wimax
Planning of Wimax
Planning of Wimax
PART I
Overview of WiMAX Cellular Technology
35
37
39
41
Table 2.1 WiMAX and WiBRO time and frequency parameters, using a1024-point FFT.
Parameter
Nominal bandwidth
Subcarrier spacing
Useful symbol time (Ts = 1/Subcarrier spacing)
Guard time (Tg = Ts/8)
OFDMA symbol duration (Ts + Tg)
Number of symbols in frame
TTG+RTG
Frame length
Sampling frequency
(Fs = FFT points subcarriers spacing)
Physical slot (PS) ($/Fs)
Mobile WiMAX
10 MHz
7 MHz
10.9375 kHz
7.8125 kHz
91.4 s
128 s
11.4 s
16 s
102.9 s
144 s
47
33
464 PS
496 PS
5 ms
5 ms
11.2MHz
8 MHz
357.14 ns
500 ns
WiBRO
8.75 MHz
9.7656 kHz
102.4 s
12.8 s
115.2 s
42
404 PS
5 ms
10 MHz
400 ns
43
16QAM
64QAM
Code rate
1/2CTC, 6x
1/2 CTC, 4x
1/2CTC, 2x
1/2 CTC, 1x
1/2CTC
1/2CTC
3/4CTC
1/2CTC
2/3CTC
3/4CTC
5/6CTC
DL Data
rate (Mbps)
0.53
0.79
1.58
3.17
4.75
6.34
9.50
9.50
12.67
14.26
15.84
UL Data
rate (Mbps)
0.38
0.57
1.14
2.28
3.43
4.57
6.85
6.85
9.14
10.28
11.42
10 MHz Channel
DL Data
rate (Mbps)
1.06
1.58
3.17
6.34
9.50
12.07
19.01
19.01
26.34
28.51
31.68
DL Data
rate (Mbps)
0.78
1.18
2.35
4.70
7.06
9.41
14.11
14.11
18.82
21.17
23.52
Zones can be congured to make use of multi-antenna technology, including phased array beamforming, STC, and MIMO techniques. Matrix A is an
Alamouti-based transmit diversity technique, which involves taking pairs of
symbols and time-reversing each pair for transmission on a second antenna.
Matrix B uses MIMO spatial division multiplexing to increase the channel
capacity. For downlink MIMO, user data entering the BS is split into parallel
streams before being modulated onto the OFDMA subcarriers. As with the
single channel case, channel estimation pilots are interleaved with the data
subcarriers. For MIMO operation, the pilots are made unique to each transmit
antenna to allow a dual receiver to recover four sets of channel coefcients.
This is what is needed to remove the effect of the signal coupling that inevitably
occurs between transmission and reception.
The MS is initially only required to have one transmit antenna and support open loop MIMO. More advanced, closed loop MIMO operates by the
MS transmitting regular encoded messages back to the BS, which provide the
closest approximation to the channel seen by the MS. The BS then pre-codes
the MIMO signal before transmission, according to the channel state information (CSI) provided by the MS. The BS may also have the facility to control
the single transmitters from two MSs to act together to create a collaborative
MIMO signal in the UL.
Matrix A and Matrix B techniques can be applied to PUSC and AMC
zones, to be described next.
45
The IEEE 802.16e-2005 standard provides the air interface for WiMAX but
does not dene the full end-to-end WiMAX network. The WiMAX Forums
Network Working Group (NWG), is responsible for developing the end-to-end
network requirements, architecture, and protocols for WiMAX, using IEEE
802.16e-2005 as the air interface.
The WiMAX NWG has developed a network reference model to serve as
an architecture framework for WiMAX deployments and to ensure interoperability among various WiMAX equipment and operators.
The network reference model envisions unied network architecture for
supporting xed, nomadic, and mobile deployments and is based on an IP
service model. Below is simplied illustration of an IP-based WiMAX network
architecture. The overall network may be logically divided into three parts:
1. Mobile Stations (MS) used by the end user to access the network.
2. The access service network (ASN), which comprises one or more
base stations and one or more ASN gateways that form the radio
access network at the edge.
3. Connectivity service network (CSN), which provides IP connectivity and all the IP core network functions.
The network reference model developed by the WiMAX Forum NWG
denes a number of functional entities and interfaces between those entities.
Figure below shows some of the more important functional entities.
Base station (BS): The BS is responsible for providing the air
interface to the MS. Additional functions that may be part of the
BS are micromobility management functions, such as handoff triggering and tunnel establishment, radio resource management, QoS
policy enforcement, trafc classication, DHCP (Dynamic Host
Control Protocol) proxy, key management, session management,
and multicast group management.
Access service network gateway (ASN-GW): The ASN gateway
typically acts as a layer 2 trafc aggregation point within an ASN.
Additional functions that may be part of the ASN gateway include
intra-ASN location management and paging, radio resource management and admission control, caching of subscriber proles
and encryption keys, AAA client functionality, establishment and
AAA
MS
BS
MIP-HA
MS
BS
ASN
GW
AN
MS
IP Net
BS
IP Net
CSN
OSS/BSS
ASN
PSTN
GW
3GPP/
3GPP2
47
OFDM is an elegant and efcient scheme for high data rate transmission
in a non-line-of-sight or multipath radio environment
2.3.2 Sub-carrier Grouping and Sub-channels
Grouping of multiple subcarriers into sub-channels is used to improve system
performance. There are two types of subcarrier allocations to form subchannels. These are distributed and adjacent subcarrier allocations.
In distributed allocations the subcarriers are pseudo-randomly distributed
over the available bandwidth. Distributed allocation provides inter-cell interference averaging and frequency diversity in frequency-selective fading channels. Distributed allocation of subcarriers is preferred for mobile applications.
In adjacent allocations subcarriers adjacent to each other in the frequency
domain are grouped to form sub-channels. Adjacent allocation has advantages
in slowly fading channels, frequency nonselective channels and for implementing adaptive modulation and coding (AMC). Adjacent allocation is used in low
mobility and xed applications. For this the subscriber may be assigned the
sub-channel with the best frequency response.
2.3.3 Slot Allocation and Data Regions
The WiMAX PHY layer allocates slots and framing over the air. A slot is
dened as the minimum time-frequency resource that can be allocated by a
WiMAX system to a given link. Each slot consists of one sub-channel over
one, two, or three OFDM symbols. This depends on the sub-channelization
scheme used. A user data region is a contiguous series of slots assigned to
the given user. This is done by the scheduling algorithms. Data regions are
allocated to users based on quality of service requirements, demand and the
condition of the channel. Data is mapped to physical subcarriers in two steps.
a) The rst step is controlled by the scheduler. Data is mapped to one
or more logical sub-channels called slots. Slots may be grouped
and assigned to segments based on applications. This can be used
by a BS for different sectors in a cellular network.
b) In the second step, logical sub-channels are mapped to physical
subcarriers. During this process, pilot subcarriers are also assigned.
49
Fixed WiMAX
OFDM-PHY
3.5
256
192
8
56
1/32, 1/16, 1/8, 1/4
None
None
None
51
consists of nine contiguous subcarriers. A slot is one sub-channel wide and the
length changes according to the zone. The pilots in the DL change positions
periodically in a rotating pattern every fourth symbol.
DL FUSC: The DL FUSC zone uses all subcarriers to provide a high
degree of frequency diversity. The subcarriers are divided into 48 groups of
16 subcarriers and a sub-channel is formed by taking one subcarrier from each
group. For this zone a slot is dened as one sub-channel over one OFDMA
symbol. The pseudo-random distribution of data changes with each OFDMA
symbol over the length of the zone, which can be useful when attempting to
mitigate interference through the use of what is effectively a type of frequency
hopping. The pilots are regularly distributed. Their position alternates with
each symbol [7].
In the DL optional FUSC (DL OFUSC) pilot subcarriers are evenly spaced
by eight data subcarriers from each other. In the UL OPUSC the zone is identical with the UL PUSC except that it uses a tile size that is three subcarriers
wide by three symbols long [7].
TUSC1 and TUSC2: The total usage of sub-channels (TUSC1 and TUSC2)
zones is only available in the DL using AAS. They are both optional and
similar to DL PUSC and OPUSC but use a different equation for assigning
the subcarriers within the sub-channel [7].
With the exception of the DL PUSC, which is assigned after the DL preamble, all of the zones can be assigned in any order within the frame. The switching points between zone types are listed in the DL MAP. Figure 2.4 shows an
example of an OFDMA frame with several different types of zones [7]. From
Figure 2.4 the mandatory DL PUSC zone follows the preamble in the frame.
The DL sub-frame also shows a second PUSC zone, a FUSC zone, and an
AMC zone. The UL sub-frame follows the TTG and, in this example, contains
a PUSC, OPUSC and AMC [7].
2.3.5 Frame Structure
The OFDMA frame consists of a DL sub-frame and an UL sub-frame. The exible frame structure of the TDD signal consists of a movable boundary between
the DL and UL sub-frames. A short transition gap is placed between the DL
and UL sub-frames and is called transmit-receive transition gap (TTG). After
the completion of the UL sub-frame, another short gap is added between this
sub-frame and the next DL sub-frame. This gap is called the receiver-transmit
transition gap (RTG). The minimum time durations for these transition gaps
are called out in the 802.16 standard and are a function of the channel bandwidth and the OFDM symbol time. It is typical to dene these transition gaps in
terms of physical slot (PS) units. A PS is a unit of time dened as 4/(sampling
frequency).
An example of a mobile WiMAX frame is shown in Figure 2.5. This
gure shows the time-frequency relationship where the symbol time is shown
along the x-axis and the logical sub-channels along the y-axis. Logical subchannels are groupings of frequency subcarriers assigned to individual users.
The concept of sub-channels and zones will be covered later in this application
note. Figure 2.5 shows the DL and UL sub-frames separated by the TTG
and ending with the RTG. The gure also shows the relative position of the
preamble, frame control header (FCH), downlink media access protocol (DLMAP), and uplink media access protocol (UL-MAP) whose functions will be
discussed in the next section.
2.3.6 How OFDM Operates
To allocate a number of sub-channels to the sub-streams, without the need
for independent radio frequencies, OFDM uses the DFT (Discrete Fourier
53
N
/2
used
j 2fc t
j 2kf (tTg )
Ck e
(2.3)
y(t) = Re e
k=Nused /2
k=0
Fig. 2.6 Guard time should be larger than delay spread (Adapted from[8]).
(2.4)
(2.5)
and
y[n] =
L1
h[k]x[n k]L
(2.6)
k=0
Where,
x[n]L = x[n mod L]
(2.7)
(2.8)
55
(2.9)
Equation (2.9) shows that each input X[m]in the frequency domain is
extended by H [m]. It can be demonstrated that the channel is ISI free. Although
a circular convolution is deployed in this technique, the natural linear channel
produces a linear convolution. However the circular convolution is forged by
adding a cyclic prex to the transmitted data. A cyclic prex technique is when
bits at the end of the symbol are included at the beginning of an OFDM symbol
where ISI is expected. The technique is used to form a circular convolution
from a linear convolution to achieve more robustness in the system. The length
of the cyclic prex is usually equal to the guard interval or the length of the
anticipated ISI. The guard interval should have at least v samples, when the
delay spread of the channel has the duration of v + 1 samples. After adding
a cyclic prex to the original samples with the length of L, the input of the
channel is xcp = xLV xLV +1... xL1 x0 x1 , . . . , xL1 , where x0 to xL1 stands
for the original data while xLv to xL1 at the beginning of the vector shows
the cyclic prex symbols.
Based on the input of the channel the output would be ycp = h xcp , which
has the length (V + 1) + (V + L) 1 = L + 2V . However, 2v samples or
bits will be discarded at the receiver in order to eliminate interference with the
preceding and subsequent OFDM symbols. Therefore L samples remain at the
end. In [8], it is proved that the L sample corresponds to y = h x, therefore
the output y is disintegrated into a channel frequency response of H = DFT {h}
and channel frequency domain input of X = DFT {x}. Although the cyclic
prex provides a circular convolution and assists to provide an ISI free channel,
it is associated with some drawbacks. To implement that, more bandwidth
and transmission power are needed, as the v additive samples burdens the
bandwidth. The required bandwidth is extended by (L + V /L), in other words
the bandwidth will be (L + V /L) times more. Similarly the transmission
power required for transmission of the cyclic prex is 10 log10 (L + V /L)dB.
In conclusion, using the cyclic prex is equivalent to a loss of power and
data rate of L/L + v. To overcome the problem of decreasing the required
power, some recent works have been done; however more contributions and
improvement are required.
57
If > 0, sampling occurs at the later time compared to the ideal instant but
if the sampling is occurred at the earlier time, it is assumed that < 0. However, ISI happens if < 0 or > Tm Tg . According to the amount of ,
the receiver will lose some energy and that energy degradation results in
SNR( ) 2(/LTs )2 , which is the signal to noise ratio loss.
It is derived from the equation above that to minimize the timing error
should be as close as possible to zero or as small as possible in comparison
with the guard time.
2.3.7.2 Frequency Synchronization
As mentioned earlier frequency synchronization is very serious in an OFDM
system and lack of an efcient algorithm for the synchronization may result
in undesirable overlapping of the subcarriers instead of being isolated from
each other. Based on the theory, the frequency offset () should be equal to
zero to provide an interference free channel, however, the frequency offset
is unequal to zero in the practical world, which result from inconsistencies
in the oscillators at the transmitter and receiver and also Doppler frequency
shift due to mobility. Using crystal oscillator is very costly, therefore systems
must tolerate to some extent frequency offsets. The formula presented below
expresses that as the frequency offset increases, the SNR decreases:
SNR =
x /N0
= 1 + C0 (L TS )2 SNR
x /(N0 + C0 (L TS )2 )
(2.10)
59
probability density function (PDF) of the k th users channel gain is dened as:
2
2hk ehk if hk 0
(2.11)
p(hk ) =
0
if hk < 0
Based on the assumption that the base station transmits to the user with
the highest channel gain (hmax ):
p(hmax ) = 2khmax (1 ehmax )k1 ehmax
2
(2.12)
The equation exhibits that increasing the number of users results in more
probability of getting a large channel and consequently enhancing the capacity
also BER (Bit Error Rate).
2.4.3 Adaptive Modulation
Adaptive modulation refers to the selection of different modulation and coding
schemes by the transmitter and allows for the scheme to change on a burstby-burst basis per link, based on the state of the channel [8]. In other words, a
transmitter has to transmit the data in as high as possible data rate if the channel
has a good and proper condition. However, if the condition of the channel is
weak, the data rate must be as low as possible. The condition of the channel
is mostly determined through the SINR measurement. To perform that, the
transmitter has to know the channel information through the receiver feedback.
The SINR at the receiver is attained by multiplying the transmitter power and
SINR at the transmitter. According to the SINR, the best conguration is
selected. For example if the SINR is high, larger constellation and higher
error correcting rate should be deployed for achieving the higher data rate;
Whereas, if the SINR is low, to maintain a lower data rate, lower constellation
and error correcting rate has to be chosen.
In the OFDMA systems, based on the amount of the SINR and its variety
among users, different blocks of subcarriers are allocated to the users. Each
block provides the best conguration for a specic user or group of users. Different congurations are called burst proles [8]. Table 2.4 displays different
burst proles in WiMAX.
The different modulation schemes employed in WiMAX are shown in
Table 2.5.
3.5 MHz
1.25 MHz
5 MHz
256 OFDM
128 OFDM
512 OFDM
8/7
28/25
28/25
PHYLayer Data Rate (Kbps)
DL
UL
946
1,882
2,822
3,763
5,645
5,645
7,526
8,467
9,408
326
653
979
1,306
1,958
1,958
2,611
2,938
3,264
DL
504
756
1,008
1,512
1,512
2,016
2,268
2,520
UL
154
230
307
461
461
614
691
768
DL
UL
Not applicable
2,520
653
3,780
979
5,040
1,306
7,560
1,958
7,560
1,958
10,080
2,611
11,340
2,938
12,600
3,264
10 MHz
1024 OFDM
28/25
DL
UL
5,040
7,560
10,080
15,120
15,120
20,160
22,680
25,200
1,344
2,016
2,688
4,032
4,032
5,376
6,048
6,720
Downlink
BPSK, QPSK, 16 QAM, 64 QAM;
BPSK optional for OFDMA-PHY
Mandatory: convolutional codes at
rate 1/2, 2/3, 3/4, 5/6
Optional: convolutional turbo codes
at rate 1/2, 2/3, 3/4, 5/6; repetition
codes at rate 1/2, 1/3, 1/6, LDPC,
RS-Codes for OFDM-PHY
Uplink
BPSK, QPSK, 16 QAM; 64 QAM
optional
Mandatory: convolutional codes at
rate 1/2, 2/3, 3/4, 5/6
Optional: convolutional turbo codes
at rate 1/2, 2/3, 3/4, 5/6; repetition
codes at rate 1/2, 1/3, 1/6, LDPC
61
by changing the FFT size based on the bandwidth, which can underpin the
NLOS operations. OSFDMA also supports for MIMO technique, Advanced
Modulation and Coding and some other important features.
Adaptive antenna gain is n times more than the gain of individual element
of the antenna array, if each array consists of n individual antennas, with the
equal antenna gains [14].
G(AdaptiveAntenna) = nG(AntennaElements)
(2.13)
63
path between the transmitter and receiver [15,16]. However, achieving a perfect diversity is impossible in practical environments. To achieve a desirable
diversity, isolation of the antenna elements is crucial. Since the mutual coupling intensies the correlation between antenna ports, the efciency of the
diversity technique drops because for achieving a high diversity gain, correlation between the ports needs to be minimized. Maintaining the orthogonality
of the antenna elements as much as possible assists in the achievement of more
isolation. However, according to Chu-Harrington [16], to maintain the orthogonality, the size of the antenna needs to be enlarged. Therefore, in the concept
of the MIMO antenna technology, one of the most important challenges is to
implement the multiple antennas in an ever-decreasing space [16].
References
[1] Yan Zhang. WiMAX Network Planning and Optimization. USA: CRC Press, 2009.
[2] Johnson I Agbinya. IP Communications and Services for NGN. New York: Taylor and
Francis, 2009.
[3] Kejie Lu, Yi Qian, Hsiao-Hwa Chen, Shengli Fu. WiMAx Networks: From Access to
Service Platform. IEEE Computer Society vol 22, pp. 3845, May/June 2008.
[4] Zakhia Abichar, Yanlin Peng, J. Morris Chang. WiMAX: The Emergence of Wireless
Broadband. IEEE Computer Society, vol. 8., pp. 4448, July/August 2006.
[5] Kosta Tsagkaris, Panagiotis Demestichas. WiMAX Network. Vehicular Technology
Magazine, IEEE. Vol. 4, pp: 2435, June 2009.
[6] Eklund, Carl, et.al. WirelessMAN, Inside the IEEE802.16 Standard for Wireless
Metropolitan Networks, IEEE Press, 2006.
[7] Peter Cain, Mobile WiMAX PHY (RF) operation and measurement, Part 1.
[8] Jeffrey G. Andrews, Arunabha Ghosh, Rias Muhamed. Fundamentals of WiMAX, USA:
Prentice Hall, 2007.
[9] Hujun Yin, Siavash Alamouti. OFDMA: A Broadband Wireless Accss Technology.
Sarnoff Symposim. pp: 14, 2006.
References
65