2. Multiple Access Technology
User Frequency
Time
Power
Traffic channels: different users
are assigned unique code and
transmitted over the same
frequency band at the same
time, for example, WCDMA and
CDMA2000
Traffic channels: different frequency
bands are allocated to different
users,for example, AMPS and TACS
Traffic channels: different time slots
are allocated to different users, for
example, DAMPS and GSM
Frequency
Time
Power
Frequency
Time
Power
FDMA
TDMA
CDMA
User
User
User
User
User
3. Duplex Spacing: 190 MHz
FDD
Time
Frequency
Power
5 MHz 5 MHz
Code Multiplex
UL DL
UMTS USER 1
UMTS USER 2
Time
Frequency
Power
TDD
5 MHz
Code Multiplex
&
Time Division
666.67 µs
DL
UL
DL
DL
UL
UMTS USER 2
UMTS USER 1
Multiple Access Technology
• WCDMA: FDD or TDD
Uplink: 1920 MHz - 1980 MHz
Downlink: 2110 MHz - 2170 MHz
Each carrier is 5 MHz width.
4. Binary data to transmit 0 1 0 0 1 0Binary data to transmit 0 1 0 0 1 0
The faster is the bit rate, the more the energy is spread on the spectrumThe faster is the bit rate, the more the energy is spread on the spectrum
+ a
- a
a2
T0
s(t)
T0
1/T0 2/T0
Frequency
Time
0 1 0 0 1 0
+ a
- a
a2
T1
s(t)
T1
1/T1 2/T1
Frequency
NRZ
coding
Time
0 1 0 0 1 0
Power
spectrum
Spread Spectrum Principle
• 1 - Time - Frequency Duality
5. Tbit
Tchip
Data sequence
spreading sequence
transmitted sequence
a2
Tbit = Ebit
1/Tbit
Tchip = Echip
1/Tchip
Frequency
a2
Tchip
1/Tchip
+a
-a
-1
+1
-a
+a
x
=
Data
sequence
Transmitted
signal
Spreading sequence generator
Modulation
x(t)
Power spectrum
Spread Spectrum Principle
• 2 – Transmission (Spreading)
6. Tbit
Tchip
Data sequence
spreading sequence
received sequence
a2
Tbit = Ebit
Power spectrum
1/Tbit
Tchip = Echip
1/Tchip
Frequency
a2
Tchip+a
-a
-1
+1
-a
+a
x
=
1/Tchip
Received
signal
Data
sequence
Spreading sequence generator
Demodulation
x(t)
Spread Spectrum Principle
• 3 – Reception (Dispreading)
11. WCDMA Principles
• Multiplexing users data
Power spectrum
User 1
User 2
User 3
User 4
User 5
Spreading
Code 1
Code 2
Code 3
Code 4
Code 5
Composite signal
5 MHz
Codes discriminate users
15. Scrambling codeScrambling code
Channelization code 1Channelization code 1
Channelization code 2Channelization code 2
Channelization code 3Channelization code 3
User 1 signal
User 2 signal
User 3 signal
BTS
Code Multiplexing
• 1 - Downlink Transmission on a Cell Level
16. BTS
Scrambling code 3
User 3 signal
Channelization code
Scrambling code 2
User 2 signal
Channelization code
Scrambling code 1
User 1 signal
Channelization code
Code Multiplexing
• 2 - Uplink Transmission on a Cell Level
17. Functions and Features of the Scrambling and
Channelization Codes
Channelization (Orthogonal) code Scrambling (Pseudorandom) code
Purpose
Uplinks: Distinguish physical data
(DPDCH) and control channels
(DPCCH).
Down links: Distinguish the down
links of different users in the same
cell.
Uplinks: Distinguish terminals
Down links: Distinguish cells
Length
4~256 chips (1.0-66.7 us)
The down links contain 512 chips
Uplinks: 10 ms = 38400 chips or
=66.7 us = 256 chips
Down links: 10 ms = 38400 chips
Code cluster
OVSF (Orthogonal Variable Spreading
Factor)
Long 10 ms code: Gold code
Spreading
spectrum
Yes, transport bandwidth is added.
No, transport bandwidth is not
affected.
21. Interference level
Example: 2 UEs at the
same distance from the
BTS using 2 data rates
Eb/No
require
d
SF=128
Service provided: Speech
Interference level
Eb/No
require
d
SF=8
Service provided: Data 144
User 2 needs more power for the
UL & DL for the same quality as
user 1
UE2
UE1
Speech 8 kbps Data 144 kbps
The higher the SF, the less power requiredThe higher the SF, the less power required
BTS
Received power
Received power
Coverage Limits (1)
22. SF = 128
Speech 8 kbps Data 64 kbps Data 384 kbps
BTS
SF = 32
SF = 4
Coverage Limits (2)
23. Receiver sensitivity (x kbps)
BS Receiver
Maximum Noise Floor
Lowest Despread Signal
BTS
UE1
x kbps x kbps
UE2 UE3
x kbps
Eb/No
Processing
Gain
Uplink Limits (1)
The transmission medium is a resource that can be subdivided into individual channels according to different criteria that depend on the technology used. Here’s how the three most popular radio technologies establish channels: FDMA (Frequency Division Multiple Access) each user is on a different frequency A channel is a frequency. TDMA (Time Division Multiple Access) each user is on a different window in time (“time slot”) A channel is a specific time - slot on a specific frequency. WCDMA (Wide-band Code Division Multiple Access) Each user uses the same frequency all the time, but it is mixed with different distinguishing code patterns. A channel is a unique (set of) code pattern(s).
The possibility to operate in either FDD or TDD mode is allowed for efficient utilization of the available spectrum according to the frequency allocation in different regions. FDD and TDD are defined as follows: FDD A duplex method whereby the u plink and d ownlink transmissions use 2 separate frequency bands: Uplink : 1920 MHz - 1980 MHz Downlink : 2110 MHz - 2170 MHz Each carrier is 5 MHz wid e and the u plink channel is 190 MHz away from the d ownlink. So there are up to 12 pairs of carriers. TDD A duplex method whereby the u plink and d ownlink transmissions are carried over same frequency using synchronized time intervals. The carrier still use s a 5 MHz band. FDD mode is the preferred mode for macro-cellular applications. TDD mode is the preferred mode for the unpaired part of the spectrum. Because each time - slot can be assigned a different direction, the TDD mode offers a great flexibility to manage duplex and asymmetric traffic. The TDD spectrum will be used for low mobility coverage in urban areas.
The information to transmit is a succession of bits, i.e. “0” or “1”. Bits, unlike electrical signal s have no physical existence. It is the purpose of data modulation (example above: NRZ used in UMTS) to give a physical existence to the virtual bits. When looking at the power spectrum of a succession 0 and 1 coded with NRZ modulation, we can see that the spectrum is composed of lobes that cut the frequency axis at multiple s of the bit period T. The main part of the power (90%) is located in the first lobe which has a maximum value for f = 0 of a 2 T = Eb. Nevertheless, the remaining parts are consuming spectrum as the theoretical spectrum is infinite. So, the faster the modulation is (the smaller T is ), the more the energy is spread o ver the frequency domain.
Spreading the spectrum consists in artificially increasing the modulation rate (chip rate) in order to spread the energy of the information signal on a wide frequency band without modifying the data rate. The number of chips per bit is called the Spreading Factor (SF) and defines the data service required for the user: For UMTS: Bit Rate x SF = 3.84 Mchip/s (Chip Rate) The following table shows examples of data services and associated Spreading Factors:
To be able to perform the de - spreading operation, the receiver must not only know the sequence used to spread the data signal, but the spreading sequence of the received signal and the locally - generated spreading sequence must be synchronized. This synchronization must be accomplished at the beginning of reception and maintained until the whole signal has been received.
To be able to perform the de - spreading operation, the receiver must not only know the sequence used to spread the data signal, but the spreading sequence of the received signal and the locally - generated spreading sequence must be synchronized. This synchronization must be accomplished at the beginning of reception and maintained until the whole signal has been received.
To be able to perform the de - spreading operation, the receiver must not only know the sequence used to spread the data signal, but the spreading sequence of the received signal and the locally - generated spreading sequence must be synchronized. This synchronization must be accomplished at the beginning of reception and maintained until the whole signal has been received.
To be able to perform the de - spreading operation, the receiver must not only know the sequence used to spread the data signal, but the spreading sequence of the received signal and the locally - generated spreading sequence must be synchronized. This synchronization must be accomplished at the beginning of reception and maintained until the whole signal has been received.
In a multi-path environment, the original transmitted signal is reflected by obstacles such as buildings or mountains, and the receiver has to treat several copies of the signal with different delays. Actually, from each multi-path point of view, other multi- path signals are considered as interference and are partially suppressed after de - spreading thanks to the processing gain. However, a further benefit is obtained if several multi-path signals are combined together using a rake receiver. The rake receiver has multiple fingers (4 to 8), each for a multi-path component. In each finger, the received signal is de - spread by the code which is time aligned with the delay of the multi-path signal. After despreading, the signals are combined using either equal gain or maximum ratio combining. This technique provides a more stable transmission channel. Rake receivers are used in the uplink and the downlink. In addition to multi-path combination the rake receiver is used by the UE to communicate with several cells (macro - diversity) . A trade-off has to be made between the multi-path gain and the capacity loss due to the use of multiple channels. Another receiver, which has been under study for a certain time now, works in a totally different way. MUD (Multi-User Detection) removes the unwanted multiple access signals through a complex algorithm. Its goal is to cancel the intra-cell interference. By so doing, an increased capacity and coverage are expected. Also, this would cancel the near-far problem, but power control would still be needed to limit inter-cell interference. This receiver is a bit more complex than the rake receiver.
All WCDMA users occupy the same frequency at the same time. Frequency and time are not used as discriminators. WCDMA operates by using CODES to discriminate between users. The receiver will ‘hear’ all the transmitter signals mixed together . B ut by using the correct code sequence , it can decipher the required transmission channel and the rest is background noise. Spreading sequences are actually unique streams of 1 and -1 which compose the code associated with a user. Therefore, users are discriminated thanks to spreading codes . Many code channels are individually “spread” with their associated “code” and then added together to create a “composite signal”. In the receiver, the composite signal is correlated with a replica of the code used to spread the data to be recover ed . Thus, low cross-correlation between the desired users and the interfering users is important to suppress multiple access interference. Good auto-correlation properties are required for initial synchronization and to reliably separate multi-path components. Remark : The correlation between two bit strings of the same length is defined as the “degree of similarity” between them: When the correlation is determined between two copies of the same string, it is called auto-correlation . When the correlation is determined between any two same length strings, it is called cross-correlation .
User A after spreading spreading sequence = +1 -1 +1 -1 = +a -a +a -a User B after spreading spreading sequence = +1 -1 -1 +1 = +b -b -b +b The signals will be added together and received as: (+a +b) (-a -b) (+a -b) (-a +b) The r eceiver reapplies the sequence spreading sequence = +1 -1 +1 -1 therefore +1x(+a +b) = +a +b -1x (-a -b) = +a +b +1x(+a -b) = +a -b -1x(-a +b) = +a -b therefore +a +a +a +a = 4a +b +b -b -b = 0 100% of the results are ‘a’ and 0% are ‘b’, so we assume ‘a’. If we apply the spreading code for user B to the same received signal we will receive a result in favor of ‘b’.
At the receiver, as the codes are different and are known, only the power of the intended user is de-spread. After despreading (decoding), correct data recovery requires a given value for the Eb to No ratio. Under this Eb/No ratio the noise will generate too many errors. The noise is mainly generated by the other users transmitting at the same time and at the same frequency but using different spreading codes. Therefore, in order not to cross this maximal noise level, all the users have to share their power: In WCDMA the Time-Frequency plane is not divided among the mobile subscribers as is done in TDMA or FDMA. So the common shared resource is power. The de-spreading process gives processing gain proportional to the bandwidth of the spreading signal. The larger the s preading f actor, the larger the gain. This means that by using a larger s preading f actor, we can reduce the power (and therefore the background noise). Thanks to this property, spread signals can operate at negative signal - to - noise ratios provided that they possess enough gain. Example: The narrow-band signal requires an Eb/No of 12 dB to achieve a certain bit - error rate performance. What is the required Ec/No, knowing that the processing gain is 20 dB?
WCDMA interference come s mainly from nearby users. Transmit power on all users must be tightly controlled so their signals reach the base station at the same signal level. This way, interferences are controlled and the famous near-far problem is alleviated. Power control is also done in the downlink to decrease the inter-cell interference. The Eb/No target is set for every service, and for each environment. Every constructor tries to have the lowest Eb/No target possible. For example, it could be worth 6.1 dB for 12.2 kbps speech in the downlink, in a dense urban area.
A mobile station or UE is surrounded by BSs ( Base Station ) , all of which transmitting on the same WCDMA frequency. It must be able to discriminate between the different cells of different b ase s tations and listen to only one set of code channels. Therefore two types of codes are used: Channelization The user data are spread synchronously with different channelization codes. The orthogonality properties of OVSF enable the UE to recover each of its bits without being disturbed by other user channels. Scrambling S crambling is used for b ase s tation and UE identification. It reduces the interference with neighboring cells since the same channelization codes are used . It is important to maintain good cross correlation characteristics between the different scrambling codes in order not to decode an interferer. O nce allocated to the U E s , t hese codes, remain the same during the whole communication. Otherwise, the b ase s tation must be notified of the change. Similar to the re-use of frequency in GSM, scrambling codes are necessarily re-used.
The WCDMA system must be able to uniquely identify each UE that may attempt to communicate with a B S . The different uplink code channels are distinguished by different UE scrambling codes. They may be scrambled by either long or short scrambling codes. Orthogonality from channelization codes is lost because there is no longer synchronous transmission. Nevertheless UE scrambling codes reintroduce some kind of reduced orthogonality thanks to their good cross-correlation properties.
In order to satisfy the request of UE4, UE1 is handed over to another cell if th is is possible. If not, the access to UE4 could be denied.
In order to satisfy the request of UE4, UE1 is handed over to another cell if th is is possible. If not, the access to UE4 could be denied.
The channelization codes are OVSF ( Orthogonal Variable Spreading Factor ) codes that preserves the orthogonality between a user’s different physical channels. The OVSF codes can be defined using a code tree. In the code tree the channelization codes are uniquely described as C ch , SF , k , where SF is the Spreading Factor of the code and k the code number, 0 k SF-1. A channelization sequence codes one user bit. As the chip rate is constant, the different length s of code enable different user data rates to be coded . The length of an OVSF code is an even number of chips and the number of codes is equal to the number of chips. The codes generated within the same layer constitute a set of orthogonal codes. Furthermore, any two codes of different layers are orthogonal except when one of the two codes is a father code of the other. For example C 4 , 3 is not orthogonal with C 1 , 0 and C 2 , 1 , but is orthogonal with C 2 , 0 . Each s ector in each b ase s tation is transmitting WCDMA d ownlink t raffic c hannels with up to 512 code channels. Exercise: Find code C ch , 8 , 3 and code C ch , 16 , 15 OVSF shortage Scrambling enables neighboring cells to use the same channelization codes. This allows the system to use a maximum of 512 OVSF in each cell. Notice that the use of an OVSF forbid the use of the other codes of its branch. This considerably reduces the number of available codes especially for fast rate service. This may lead to OVSF shortage. For such a case, secondary scrambling codes are allocated to cells and enable to re-use the same OVSF in the same cell.
Orthogonality means that there is no correlation between codes, so, C k presence does not affect C j energy. OVSF codes are completely orthogonal for zero delay. For other delay s they have very bad cross-correlation properties, and thus they are suitable only for synchronous applications. If the synchronization at To is not respected then there is no orthogonality anymore ==> C j and C k interfere.
For a given noise level, as the processing gain is smaller for a high rate user data, the acceptable path loss is lower and therefore so is the range of the cell.
The radius of a cell varies with the SF ( Spreading Factor ) and, as we will see, the noise level or, as a matter of fact, the number of active subscribers in the cell. The figures are given for a traffic load of 50% of the maximum traffic acceptable in the cell.
Consider UE 1 transmitting at the boarder of the cell. That is to say UE 1 transmits at full power and is received at the minimum power to access the cell (equal to the receiver sensitivity). After de - spreading, decoding UE 1 needs a noise level lower than the maximum noise floor fixed by the processing gain and the Eb/No. For so long as the interference generated by users UE 2 and UE 3 do es not cross this floor, UE 1 is correctly decoded.
Consider a BTS transmitting with UE1, UE2 and UE3. The further the UE is, the more power is needed from the BTS to be able to reach it. When UE4 asks for an access, the BTS doesn’t have not enough power capacity to add the power intended for UE4 although they are very close to each other. In th is case, the UE can be handed over to another BTS (if possible), or the system can degrade the quality of communication for the other UEs in order for the BTS to be able to reach it.
In order to satisfy the request of UE4, UE1 is handed over to another cell if th is is possible. If not, the access to UE4 could be denied.
In order to satisfy the request of UE4, UE1 is handed over to another cell if th is is possible. If not, the access to UE4 could be denied.
In order to satisfy the request of UE4, UE1 is handed over to another cell if th is is possible. If not, the access to UE4 could be denied.
In order to satisfy the request of UE4, UE1 is handed over to another cell if th is is possible. If not, the access to UE4 could be denied.
