Multiple Acces and Cellular Structure

7/30/2019 Multiple Acces and Cellular Structure

1/65

1

Review Questions

What are the main differentiating features of 1G,2G and 3G?

What was the bandwidth used in one-direction in1G AMPS?

In DAMPS, how many users share the samebandwidth as AMPS?

What are the range of data rates for wideband and

broadband wireless systems? What is an unlicensed band? What are some of the

typical applications?


2/65


3/65

Multiple Access Techniques


4/65

4

Multiple Access (MA)

MA is setting up as many transmission

paths as required.

MA is the way in which the common

resources (power and bandwidth) are shared

by many users.


5/65

5

Multiple Access TechniquesIn terms of circuit assignment

Pre-assignment

Demand assignment

Random assignment

In terms of modulation (waveform assignment)

Frequency division multiple access (FDMA)

Time division multiple access (TDMA)

Code division multiple access (CDMA)


6/65

6

Consider that the following samples of 3users data to be multiple accessed

User 2 data

User 3 data

User 1 data


7/65

7

Frequency Division Multiple Access

FDMA

f

bandwidth

BW

guard band


8/65

8

Waveforms of Users 1,2 & 3 after FDMA


9/65

9

Input to the amplifier after 3 FDMA

signals are added


10/65

10

Nonlinear Effects in FDMA

Received signal is sum of multiple carriers.

Receiver power amplifiers are operated

nonlinearly (near saturation) for maximumefficiency.

The nonlinearities cause intermodulation(IM) frequencies to appear in the amplifier

output. IM components can interfere with other

channels in the FDMA system.


11/65

11

Time Division Multiple Access

TDMA systems divide the radio spectrum

into time slots.

Only one user can transmit or receive

during one time slot.

Usually, each user may occupy the channel

once during a time frame, where one framecomprisesNtime slots.


12/65

12

TDMA


13/65

13

User 1,2&3 data before and after TDMA


14/65

14

TDMA

t

Slot 1 Slot 2 Slot 3 SlotN

1 frame

sync. bits

t

trail bits info bits Guard time

Each slot requires overhead bits.

More overhead reduces efficiency.


15/65

15

TDMA Systems

TDMA systems transmit data in a buffer

and burst method.

The transmission is noncontinuous.

Unlike FDMA systems which can transmit

analog signals, TDMA must transmit data

and digital modulation must be used.


16/65

16

TDMA Features

Only one carrier. No IM.

Number of time slots per frame dependson bandwidth, desired date rate,modulation technique.

Receiver must synchronize to each time slot,thus more synchronization bits are required

in TDMA compared to FDMA. It is possible to allocate more than one time

slot per framebandwidth on demand.


17/65

17

CDMA

In CDMA signals of different users overlap in time and

frequency. The separation is achieved by assigning

different codes to each user.


18/65

18

Direct Sequence Spread

Spectrum

Modulator Channel Demodulator

PN sequencegenerator

PN sequencegenerator

18181818School of Inform at ion Technolog y and Engineer ingUniversi ty of Ottawa


19/65

19

Data

PN-1

Data spread by PN-1

PN-2

Data despread by PN-2

Data despread by PN-1

1

1001010

1001010

1011100

1101001

1


20/65

20

Power Spectrum of DS-SS Signals at Tx

t f

t

f

noise level

20202020School of Inform at ion Technolog y and Engineer ingUniversi ty of Ottawa


21/65

21

Reuse Patterns

TDMA or FDMA CDMA


22/65

22

X X XB1 B2 B3

d1

d2d3

u1

u2

Interference comes from non-adjacent channels. Suppose B1 tx at

power pa, B3 tx at power pb. Let the distances from B1 and B3 to u1 be

d1 and d3 respectively.

FDMA/TDMA Forward Link C/I at u1

C = pd

I pd

CI

pp

dd

a

n

b

n

a

b

n

n

1

3

3 3

3

1bg bg gbg


23/65

23

X X XB1 B2 B3

d1

d2

d3u1

u2

Suppose u1 tx at power p1, u2 tx at power p2. Let their distance to B1

be d1 and d2 respectively.

FDMA/TDMA Return Link C/I at B1

C = pd

I pd

CI

pp

dd

n n

n

n

1

1

22

2 2

1

2

2

1bg b g gbg


24/65

24

X X XB1 B2 B3

d1d4 d3d5 d2

u1u2 u3

C =

p

d I

p

d I

p

d

C

I I

p

d

p

d

p

d

b

n

a

n

c

n

b

n

a

n

b

n2

1

5

3

4 1 3

2

5 4

bg bg bgbg

bg bg, , ,

CDMA Forward Link C/I at u2

Suppose B1 tx at power pa, B2 tx at power pb and B3 tx at power pc

Let their distances to U2 be d5, d2 and d4 respectively.


25/65

25

X X XB1 B2 B3

d1

d4d3

d5 d2u1 u2 u3

C = pd

I pd

I pd

CI I

p

dp

d

p

d

n n n

n

n n

2

2

1

1

5

3

3

4 1 3

2

2

1

5

3

4

bg bg bg bgbg bg

, , ,

CDMA Return Link C/I at B2

Suppose u1 tx at power p1, u2 tx at power p2 and u3 tx at power p3

Let their distances to B2 be d5, d2 and d4 respectively.


26/65

26

Near-Far Effect

A major problem with DS-CDMA is the

dominance of a stronger user.

Consider 4th

power propagation loss. If both usersare transmitting at the same power, what is the

received power level difference (in dB) if one user

is 10 times closer than the other?

Near-far effect demonstrated in the above example

is not critical in frequency hopped systems.


27/65

Intracell and Intercell Interference in CDMA


28/65

28

A Summary on Multiple Access

FDMA:+ Flexible (modulation, transmission rate,etc).

Inefficient use of available power to avoidintermodulation (IM) products.

Many receivers needed to receive simultaneouslyfrom several transmitters.TDMA:+ No IM, efficient resource utilization.

Synchronization is critical.

Peak to average power ratio is high.CDMA:+ Flexible (for networking)

Generally self interference limited.


29/65

29

Do systems fall into only one

of 3 categories?

f

t

f ,t

Answer: In practice NO.


30/65

Cellular Concepts


31/65

31

Clusters

Cluster size

N= 7

1

123

4

56

7

23

4

56

7

12

34

5 6 712

34

56

7

12

34

56

7

In TDMA and FDMAcells are grouped intoclusters.

The total number ofchannels are equallydistributed among allcells in a cluster.

The channelallocation pattern isrepeated in everycluster.

Cells that areallocated the samefrequencies causeinterference to oneanothercochannel

interference (CCI).


32/65

32

Clusters (2)

12

34

1

2 3 4 12

34

12

3

4

43

21

4

3 2 1

Cluster size

N= 4


33/65

33

Real Cellular Structure


34/65

34

Clusters (3)

A cluster must be more or less symmetrical

and similar North/South and East/West

dimensions.

We cannot select cluster size arbitrarily.

Cluster sizes must satisfy the following:

N= i2

+ij+j2

where i andj are nonnegativeintegers.

Thus N= 1, 3, 4, 7, 9, 11, 12


35/65

35

Frequency Re-use

The number of effective channels in a service area

can be increased by decreasing the cluster size.

However, we cannot decrease the cluster size to aslow as we like.

Decreasing cluster size decreases distance between

cells employing same frequencies which in turn

increases co-channel interference.


36/65

36

Channel Assignment Strategies

Fixed assignment strategies.

Cells are allocated a predetermined set of voice

channels.

Dynamic assignment strategies.

Voice channels are not permanently allocated to

a cell.


37/65

37

Fixed Channel Assignment

Strategies Cells are allocated a fixed set of voice channels.

Some channels are reserved for handoffs.

If all voice channels are in use, any new callrequests are denied access (blocked).

One variation allows channel borrowing. When acell has all of its channels in use, it may requestadditional channels from a neighboring cell.

The MSC supervises the borrowing procedure. TheMSC ensures that borrowing does not disrupt orinterfere with any calls in progress in the donor cell.


38/65

38

Dynamic Channel Assignment

Each time a call request is made, the BS

requests a channel from the MSC.

The MSC then allocates a channel to thecell following an algorithm that takes many

factors into account.

MSC allocates a frequency that is not currentlyin use within the minimum distance.

MSC takes likelihood of future blocking in cell.


39/65

39

Handoffs

Mobile unit periodically monitors, measures and ranks allpilots within its range

As a mobile moves from one cell (or sector) to another,connection is rerouted to the new cell.

Handoffs must be performed successfully, infrequently and

imperceptible to the users.(1) Idle hand-off(2) In-call hand-offIdle hand-offoccurs when the mobile unit is on but not

actively engaged in a call. In this state the mobile unit

monitors the paging channel. One of two conditions can trigger an idle handoff(1) Deterioration in reception quality(2) Another pilot is stronger by a certain level.


40/65

40

Handoff (handover)

Hard handoff: Channel in source cell is releasedfirst and then channel in target cell is engaged.(break before make).This is the only possibility in

TDMA or FDMA.

Soft handoff: channel in source cell is retained andused for a while in parallel with channel in targetcell.(make before break). This is both a necessity

and a blessing for CDMA. Mobile communicates with multiple BSs during

handoff


41/65

41

Prioritizing Handoffs

It is possible that many mobile users enter thesame cell, placing a heavy load on the new cell.

The rate of decrease in the received power (whichis usually a function of the velocity of the MS),determines the priority of a handoff.


42/65

42

Softer Handoff

Softer handover refers to the handover

between the sectors of a cell.

Forward link is the same as soft handover.

In return link no need to communicate with

other Base stations.Thus it is faster.


43/65

43

Advantages of Soft Handoff

Contact with a BS is made before the call is

switched, and this prevents the mobile

losing contact if handoff signal is not heard. Diversity reception provides additional

resistance to fading.


44/65

44

Practical Handoff Considerations

With frequency reuse, a cellular provider canincrease system capacity by decreasing cell sizes.

With more cells in the same area, and by using thesame frequency reuse pattern, the number ofeffective channels over the same area is increased.

However, as the cells decrease in size, the numberof handoffs increase.

More channels must be reserved for handoffs.

As cells decrease in size, high speed users may requiremany handoffs per call, which increases probability ofcall being cut off.


45/65

45

Practical Handoff Considerations

Low speed users, such as pedestrians, may not need a

handoff, even in very small cells.

How can we support both high speed and low

speed users while maintaining a high system

capacity and low number of handoffs?

Umbrella cell approach.


46/65

46

Umbrella Cell Approach

Low speed users can be handled bymicrocells or picocells.

High speed users are handled by a largercell macrocell which is co-located withsmaller cells.

Speed of MS can be estimated by MSC byobserving the rate of change of signalstrength.


47/65

47

Umbrella Cell Approach


48/65

48

Cell Dragging

Low speed users in microcells often enjoy line of

sight communications with the BS even when

outside of the boundaries of the cell.

Thus handoffs may not be made even when the

user is well within the boundaries of a new cell.

Cochannel interference to other cells will be

increased.


49/65

49

Interference and Its Effect on

Capacity Interference is the major limiting factor in the

performance of cellular radio systems.

The two major types of system-generatedinterference are co-channel interference (CCI) andadjacent channel interference (ACI)

CCI: several cells employ the same set of frequenciesin a service area. These are called co-channel cells. An

MS or the BS in a cell will receive signals from othercells which are on the same frequency.

ACI: sidelobes of signals in adjacent frequencies in thesame (or neigboring) cell can interfere with a userssignal.


50/65

50

Co-channel Interference and

System Capacity Frequencies are reused everyNcells.

A user is assigned a pair of frequencies for a fullduplex call.

The user will receive the desired signal from theBS and it will receive from all other BSstransmitting on that frequency.

It is possible that the nearest neighbours are notcurrently transmitting on that frequency.

The BS receives the signal from the MS as well asfrom other MSs in other cells transmitting on thesame frequency.


51/65

51


System Capacity 2 We cannot reduce the effect of CCI simply

by increasing the power with which we

transmit.This will create larger CCI in other cells who

will then increase their power to compensate.

The signal to CCI power ratio is determinedby the distance between co-channels and the

propagation exponent of the channel.


52/65

52


System Capacity 3 Assuming:

Equal cell radii.

All MS transmit with same output power. All BS transmit with same output power.

The worst case Signal to Interference power ratio

for an MS located on the cell boundary can be

approximated by an equation that does not dependon cell radius, but on the cluster size and the path

loss exponent.


53/65

53


System Capacity 41

2 4

56

7

1

2 3 4

56

7

12

3

4

56

7

R

(3N)1/2R

If both BSs are transmitting at same power

then signal from interfering BS is (3N)n/2

times weaker than desired signal

> (3N)1/2R


54/65

54


System Capacity 5 Assuming hexagonal cell shapes, there are 6

nearest interfering cells.

Thus C/I > (3N)n/2

/6. We use this as worst caseC/I.

Example: We are setting up a cellular service in

an area where the path loss exponent is 3.88. We

want a minimum C/I of 18 dB. What is thesmallest cluster size which accommodates this?


55/65

55


System Capacity 6 (3N)1.94/6 > 101.8.

(3N)1.94 > 63.16 = 378.6

N > 378.6(1/1.94)/3 = 7.1

N = 9

(N = 7 does not guarantee C/I > 18 dB and

N = 8 is not allowed).


56/65

56


System Capacity 7 Thus we can see that the co-channel interference

places a constraint on the cluster size which has an

effect on the system capacity.

As cluster size decreases, the number of channels per

cell increases.

If we can accept more CCI, then we can reduce the

cluster size.

Also, if the pathloss exponent is higher, then the cluster

size can also be decreased.


57/65

57

Adjacent Channel Interference

The BS receives signals from many userssimultaneously.

Although signals are filtered to reduce out-of-bandemissions, the filtering is not perfect andsidebands exist.

If all users transmit at the same power level, thenthe received power of signals originating fromvery close to the BS will be much higher than thereceived power of signals originating from nearthe cell boundary.


58/65

58

Adjacent Channel Interference 2

Signal 2s sidelobe is as strong as signal 1s mainlobe.

Performance of signal 1 is severely degraded.

Spectrum of signal 1

Spectrum of signal 2


59/65

59


Suppose a signal is received at the BS with

sidebands whose spectral density is 30 dB lower

than its main lobe.

If the adjacent signal is received 30 dB lower due

to distance, then the interfering sideband is

roughly the same strength as the desired signal.

Because of the near-far problem, a desired signalmay be lost in the sidebands of adjacent channel

signals.


60/65

60


To reduce the effect of ACI, the received powerlevels are controlled.

This is also done to ensure that the MS istransmitting at the smallest power necessary tomaintain a reliable link.

All signals should be received by the BS withpower levels that are within a few dB of oneanother.

If a signal is received with a power level above theupper limit, a control message is sent on the FCC.This instructs MS to decrease transmitted power.


61/65

61

Why Power Control?

1. To eliminate near-far effect2. To conserve transmit power.

What is power control?Power control ensures that each user receives and

transmits just enough energy to properly convey

information while interfering with other users no more

than necessary.Minimizing the transmitted power of a portable user

unit maximizes the interval between battery charges.


62/65

62

Is Power Control Critical in Both

Directions? Critical in the uplink because of Multiple

Access Interference (MAI).

In the downlink not critical for MAI but is

needed to minimize the interference into

other cells and to compensate interference

from other cells.


63/65

63

Power Control Limitations?

If power control can follow fading

perfectly, fading channel becomes an

AWGN channel. Hence the practical limitation is how

accurately and quickly can channel be

estimated. Then comes the step sizes,dynamic range etc.


64/65

64

Types of Power Controlin CDMA

In the reverse direction:

1) Open loop for coarse power adjustment (at the start)

2) Closed loop for refined adjustments (every 1.25 msec)

3) Outer loop to reduce to min level (once per frame)

In the forward direction: adjust to maintain atarget quality by:messaging, or

bit signaling


65/65

Open Loop Power Control

Adjust transmit power to compensate for distance dependentattenuation and shadowing.

Power correction based on total received power.

Mobile station transmit power defined by

tx power(dBm)+ rx power (dBm)= constant + tolerance

(typically) constant = -76 dB for 1900 MHz and -73 dB for

800MHztolerance = +/- 6dB

Tyipical time constant of open loop PC is 30 msec

Multiple Acces and Cellular Structure

Documents