CSLGSM

8/13/2019 CSLGSM

1/87

Mustafa M. Hussain 1

A PRESENTATION ON CELLULAR

NETWORKTOPICS:

1. Cellular System Overview

2. Radio Environment

3. Digital Transmission

4. Air-Interface

5. BSS in General and6. AXE based CME20 Nodes.

8/13/2019 CSLGSM

2/87


START

An introductory presentation onCellular Mobile System.

8/13/2019 CSLGSM

3/87


System Parameters: Over the air interface, speech, data and

signaling information are transmitted on amodulated bearer frequency, a carrier. Thiscarrier can be used for one or several channels.

The distance in Hz between two adjacentcarriers is called carrier spacing.

Each carrier is used in one direction only, fromterminal to base station (uplink) or base stationto terminal (downlink).

To get a duplex communication we need acarrier pair.

8/13/2019 CSLGSM

4/87


System Parameters contd..

The carriers are grouped into uplink carriers and

downlink carriers and the distance in Hzbetween the carriers, making a duplex pair, iscalled duplex spacing.

The frequency range used for carriers in one

direction is called systems bandwidth. GSM 900 has a bandwidth of 25 MHz (uplink

from 890 MHz to 915 MHz and downlink from935 MHz to 960 MHz) and a carrier spacing of

200 KHz.

Duplex carrier separation is 45 MHz.

8/13/2019 CSLGSM

5/87


System Parameters contd..

Each carrier contains 8 channels. This means8*25000000/200000=1000 channels ofsimultaneous calls.

Is that enough for a whole country or a

continent? No, and answer to the problem is frequency

reuse.

In GSM900 these carriers (frequencies) are

numbered as numerals from 1 to 124.Frequency 890.2 MHz is termed as1, 890.4 as 2and so on.

8/13/2019 CSLGSM

6/87


Frequency Reuse:

Two neighboring cells cannot use the samecarrier or frequency due to co-channelinterference.

In cellular systems, a specified number of co-

located cells form a cluster. Each frequency is only used in one of the cells

within the cluster, but reusedin thecorresponding cell in other cluster.

The cluster size (number of cells) depends onthe systems resistance to co-channelinterference.

8/13/2019 CSLGSM

7/87


Network Element:

The elements forming the mobile network have fixedpositions and are listed below-

MS, Mobile Station or mobile terminal or mobile phoneset. The MS is the user equipment and the only thing

that is mobile. BSC, Base Station Controller (used in GSM). Functions:

to control a number of BSs, adjust the output power ofpresent MSs and do handover within the BSC area.

MSC, Mobile Services Switching Center. Functions:switching, handover, subscriber services and charging.

8/13/2019 CSLGSM

8/87


Network Element contd.. VLR, Visitor Location Register. A database belonging to

each MSC, containing a list of all MSs presently being inthat MSCs service area. The list includes presentlocation area of the MS and information about subscribercategory.

GMSC, Gateway MSC is an MSC capable of connecting

a call from the fixed network to the mobile network. HLR, Home Location Register. Each GMSC has access

to this database, telling in which MSC service area theMS is at the moment.

also there are AUC (Authentication Center) and EIR(Equipment Identity Register). These are databases andare actually protection against fraud.

8/13/2019 CSLGSM

9/87


Network Element contd..

Figure-1: Cel lular Netwo rk Overview

8/13/2019 CSLGSM

10/87


Roaming:

The possibility to move around, keeping thenetwork updated about the geographicallocation, is called roaming.

An attached, idle MS is continuously measuring

the signal strength on a number of predefined(controlled channel carrying) frequencies, onefor each neighboring cell.

The MS decides which is the strongest, locks

itself to and listens on that frequency for pagingmessages and location area identity (LA-id)code.

8/13/2019 CSLGSM

11/87


Roaming contd..

Each BS continuously transmits and LA-id (LocationArea Identity) code.

As long as the MS is within the same LA nothinghappens, but if the MS reaches a cell belonging to

another LA, it will notify that by a new LA-id code andimmediately try to access the network to have its locationupdated.

If the new LA is within the same MSC service area, onlythe corresponding VLR has to be updated, otherwise the

new VLR, the HLR and the old VLR have to be updated(change of service area).

8/13/2019 CSLGSM

12/87


Paging: A call to an MS will be routed by MSC/VLR

where the MS is registered. The MSC/VLR will then send a paging message

to the MS.

This message is broadcasted all over the LA,

which means that all BSs within the LA willtransmit a paging massage to the MS.

The MS listens to one BS, hears the pagingmessage and answers it immediately.

After identification of the MS and setup of atraffic channel have taken place, the MS willinform its user by a ringing tone.

8/13/2019 CSLGSM

13/87


Handover: Changing to a new traffic channel during call set-up

or busy state is called handover.

GSM uses mobile assisted handover (MAHO),which means that the MS assist by measuring thesignal strength and transmission quality of the trafficchannel in use, and also the signal strengthneighboring base stations.

MS reports those measurement values to the BSC.

The BS is also measuring on the active channel, thismeasuring and reporting is continuously going onwhen the MS is busy.

8/13/2019 CSLGSM

14/87


Handover contd..The BSC evaluates the measurements from the MS and

BS and decides if- everything is ok

the MS has to increase or decrease its output power,due to changed distance from the BS, or

the quality is too bad and it is necessary to change basestations- make handover.

Handover includes rerouting of the call from one BS toanother and changing traffic channels.

After the handover, the MS will be informed about theneighboring cells, making signal strength measurementon them possible.

8/13/2019 CSLGSM

15/87


THE RADIO ENVIRONMENT:

8/13/2019 CSLGSM

16/87


The Radio Path:Access Methods:

The frequency spectrum is used in a multiple assessfashion, meaning that every terminal can make access to

a common resource (the frequency range) on equal

conditions.

Here are three methods:

FDMAFrequency Division Multiple Access

TDMATime Division Multiple Access

CDMACode Division Multiple Access

8/13/2019 CSLGSM

17/87


The Radio Path contd.. In FDMA, each channel corresponds to one frequency

pair. This method is used in analogue cellular systems. In TDMA each frequency is divided into time slots. Each

channel corresponds to a time slot pair. This method is

used in digital cellular systems today.

Channel separation in CDMA is achieved through theuse of different codes. All channels use the same wide

frequency range and each channel corresponds to a

unique code.

8/13/2019 CSLGSM

18/87


Frequency: The radio spectrum, 3 KHz- 300 GHz, is divided into

eight frequency bands, from VLF (Very Low Frequency)to EHF (Extremely High Frequency).

Low frequencies bounce on the earth surface and in theatmosphere and can reach very farto other countriesor continents. High frequencies, above ~ 3 GHz are

sensitive to shadowing and need line of sight (comparewith radio links).

The cellular systems use frequencies around 1 GHz inthe UHF (Ultra High Frequency) band and that is high

enough to get the right cell size and low enough to havefrequencies that can go through and bounce onbuildings.

8/13/2019 CSLGSM

19/87


Fading:

The signal strength variation of the received

signal is called fading- time, distance and

frequency dependent attenuation over the air

interface.The most important fading types are:

Path Loss Fading

Log-normal-fading (due to shadowing)

Rayleigh fading (multi-path fading)

8/13/2019 CSLGSM

20/87


Fading contd..Path Loss:

Path loss is the phenomenon which occurs whenthe received signal becomes weaker and weakerdue to increasing distance between mobile andbase station. There are no obstacles between

transmitting (Tx) and receiving (Rx) antenna. For thefree space case we say that for a given antenna thereceived power density is inversely proportional tothe square of the distance, d, between the TxandRxantennas. The received power is also said to be

inversely proportional to the square of thetransmitting frequency, f.

8/13/2019 CSLGSM

21/87


Fading contd..Log-normal-fading:

The number and size of hills,

buildings and other obstacles

between the transmitter and

receiver change when the MS

moves around. This gives riseto log-normal-fading due to

shadowing. Too heavily

shadowed areas cause a

decrement of quality. By an

appropriate cell planning,

problems with shadowing can

be solved.

Figure-2: Log -no rmal- fading

8/13/2019 CSLGSM

22/87


Fading contd..Rayleigh (Multi-path) fading:

Rayleigh fadingis a city problem and occurs when the signalgoes more than one way from the transmitter to the receiver

(due to bouncing on buildings).

Figure-3: Mult i path (Rayleigh ) fadin g

8/13/2019 CSLGSM

23/87


Multi-path fading contd..

The received signal is a sumof many identical signalswhich differ only in phase(and to some extent inamplitude). While we add

signals like vectors, it canunfortunately mean that thevector sum turns out to bevery close to zero whichmeans that the signalstrength also becomes very

close to zero - a very severefading dip indeed.

Figure-4: Ad ding two ant i-phase signals

8/13/2019 CSLGSM

24/87


Fading contd..An illustration of how the signal strength may look like at the MS

Rxantenna as you move away from the BTS Txantenna is

shown in Figure-5. It covers all the fading types discussed so far.

Figure-5: Rx sign al strength v ersus d istance

8/13/2019 CSLGSM

25/87

8/13/2019 CSLGSM

26/87


Time Dispersion:

The introduction of digital transmission brings anotherproblem: time dispersion.

This also has its origin in reflections, but in contrast tomulti-path fading, the reflected signal comes from an

object far away from the Rxantenna, say in the order ofkilometers.

The time dispersion causes Inter Symbol Interference(ISI).

ISI means that consecutive symbols interfere with eachother and it gets difficult on the receiver side to decidewhich actual symbol is detected (or actually, sent).

8/13/2019 CSLGSM

27/87


Time Dispersion contd..

An example of this is shownin Figure-6 . The sequence1, 0 is sent from the basestation. If the reflected signalarrives exactly one bit time

after the direct signal, thenthe receiver will detect a 1from the reflected wave atthe same time, as it detectsa 0 from the direct wave. Thesymbol 1 interferes with the

symbol 0.

Figure-6: Time Dispersion

8/13/2019 CSLGSM

28/87


Time Dispersion contd..

In GSM , the net bit rate over the air interface is 270.8kbit/s, which leads to a bit time of 3.7 s.

One bit therefore corresponds to 1.1 km, so if we have areflection from 1 km behind the mobile station, thereflected signal will have a 2 km longer path than the

direct one. This means that the reflection will mix a signal coming

two bittimes later than the wanted signal, with thewanted signal.

Time dispersion would appear to be a tricky problem,

and it is, but further on in this chapter we will discoverthat it might not be as bad as it seems.

8/13/2019 CSLGSM

29/87


Time alignment:

Using TDMA means that the mobile must transmit onlyduring the allocated time slot and be silent the rest of thetime. Otherwise it will interfere with calls from othermobiles using different time slots on the same carrier.Say that a mobile is very close to the base station. It isallocated time slot 3 (TS 3) and is only using this time

slot for the call. During the call, the mobile moves awayfrom the base station so whatever is sent from the basestation arrives later and later to the mobile and theanswer from the mobile also arrives later at the basestation. If nothing is done, the delay will eventually be so

long that what this mobile transmits in TS 3 will overlapwhat the base station receives in TS 4 - another call.

8/13/2019 CSLGSM

30/87


DIGITAL TRANSMISSION-

PROBLEMS AND SOLUTION

8/13/2019 CSLGSM

31/87


Transmission:

Figure-7 shows the

signal processing

blocks very

schematically. We

can see that signal

processing isperformed both in

the mobile

equipment and on

the network side.

Figure-7: Schematic Signal Process ing

8/13/2019 CSLGSM

32/87


Transmission contd..Mobile transmitting side :

First the analogue speech is digitized by the A/Dconverter (analoguetodigital).

Then it is divided into segments of 20 ms, which are fedinto the speech coder for reduction of the bit rate.

The next step is channel coding and interleaving(processes enabling error correction and error detectionon the receiving end).

Ciphering of the speech (to protect from eavesdropping),and burst formatting (adding start and stop bits, flags

etc.) is carried out. The last step is to modulate the bit stream on a carrier

and then to transmit the signal.

8/13/2019 CSLGSM

33/87


Transmission contd..

Receiving side:

On the receiving side the corresponding procedure isperformed.

The difference between the mobile equipment side and thenetwork side is mainly that the speech is not A/Dor D/Aconverted on the network side.

The network will transmit digital signals through the networkwhile the mobile equipment has to convert it tounderstandable speech directly.

If we have data instead of speech to transmit, of course noA/Dor D/Aconversion is performed on the mobile side,neither is the data fed through the speech coder.

There is also another type of channel coding, since data ismuch more sensible to transmission errors.

8/13/2019 CSLGSM

34/87


SPEECH CODING:If we look at PCMcoded speech, (see Figure 1) we see thateach speech channel delivers 64 kbit/s. 8 such channels

would then give rise to a bit rate of 512 kbit/s over the airinterface. That is with no transmission reliability added.

Figure 1: Multiplexing 32 channels on to one PCM link.

8/13/2019 CSLGSM

35/87


SPEECH CODING contd..

We must somehow lower the bit rate significantly for each

speech channel to be able to keep within the allowedfrequency band. This is accomplished by speech coding .

Since we have to lower the bit rate it is not possible totransmit the speech itself, or a digitized version. Wetransmit information about the speech instead (filterparameters and information about the excitation sequence).

As the speech organs are slow in adapting, we can say thatthe filter parameters representing the speech organs areapproximately constant for 20 ms.

Since we know how speech is created - we can create allparametric model of the speech. Speech coders are such

equipments.

8/13/2019 CSLGSM

36/87



There are mainly

three types of

speech coders,

these are-

Waveform coders.

Vocoders.

Hybrid coders.

Figure 2: Speech Quality vs. Bit Rate

8/13/2019 CSLGSM

37/87



In public mobile telephony the requirements on speech quality are

between good and excellent in figure 2. It is difficult to distinctlydefine speech quality, but when speaking about radio, three grades ofquality can be defined.

Grade 1, Telephony quality:

This quality corresponds to good connections in the public switchedtelephony networks. A standard speech coder (64 kbit/s PCM) can

handle this, even with a small margin. The usual requirements onspeech quality in mobile telephony are the same or slightly lower.

Grade 2, Communication or mobile radio quality:

Compared to grade 1 we have a fully noticeable background noises.

Grade 3, Vocoder or military radio quality:

The primary requirement is a good ability to understand what is said. As

long as this requirement is fulfilled a good portion of distortions of thespeech is tolerated.

8/13/2019 CSLGSM

38/87



Waveform coder:

If relatively high bit rate can be accepted good qualitycan be achieved with simple speech coders of thewaveform type. The 64 kbit/s PCM coder, which is astandard within the fixed telephony network, is of thistype.

The required bit rate can be reduced if we use thecorrelation between sequential samples of the speech.

The most advanced waveform coders can with a bit rateas low as 16 kbit/s produce a speech of nearly the samequality as the 64 kbit/s PCM. If the bit rate is reducedbelow 16 kbit/s the speech quality decreases quickly.

8/13/2019 CSLGSM

39/87



Vocoder:

The vocoder principle is based on a very simplified model ofthe speech generation process. This makes it possible to usea low bit rate, but the simplified model does not consider thefine structure of the speech, which severely affects thesubjective speech quality.

The reproduced speech has a synthetic and metallic tone.

It can also be difficult to recognize the speaker, although theability to understand what is said is quite good.

The vocoders are typically designed for bit rates of 1.2 or 2.4kbit/s.

Increase of the bit rate over 4.8 kbit/s hardly increases the

speech quality.

8/13/2019 CSLGSM

40/87

8/13/2019 CSLGSM

41/87



Hybrid coders:

Depending on their importance in describing the speechthe bits out from the speech coder will be divided into 3groups, very important, important and not so important

bits, and differently channel-coded. The difficulties in maintaining a fully acceptable speech

quality, including a reasonable stability with regards tovariations in the input speech signal (difference betweenmale and female voices, acoustic background noise,

etc.) increases very quickly when further reductions ofthe bit rate are introduced.

8/13/2019 CSLGSM

42/87


Channel coding:

With digital transmission, the quality of the transmitted signalis often expressed in terms of how many of the received bitsare correct leading to the expression Bit Error Rate (BER).

BER says how many percent of the total number of bits arewrongly detected. Of course we want this number or rate tobe as small as possible.

It is, however, not possible to reduce it to zero, due to theconstantly changing transmission path. This means that wehave to allow a certain amount of errors and yet be able torestore the information, or at least be able to detect errors sothat we do not use the information as if it were true. This isespecially important when we send data - we can accept

lower quality for speech.

8/13/2019 CSLGSM

43/87

8/13/2019 CSLGSM

44/87


Channel coding contd..In convolution coding, the block of coded digits generated by thecoder depends not only on the digits in the current message blockwhich is shifted into the coder but also depend on bits inpreceding message blocks. Following figure illustrates theprinciple.

Figure 4: Convolution coding.

For each new bit fed into the convolution coder the output

will be two bits. The rate is said to be 1:2.

8/13/2019 CSLGSM

45/87


Channel coding contd.. Block codes are often used when we have block-orientated

signaling, such as in analogue land mobile radio where thedata is sent in blocks. It is often used to detect errors whenARQ is implemented (Automatic Repeat Request ). In ARQwhen we detect an error, we ask for re-transmission.

Convolution coding is more associated with error correction,when for example the ARQ facility is not available. When we

digitize speech and transmit it, we cannot retransmit as thiswould lead to intolerable delays.

In GSM, both methods are used. First, some of theinformation bits are block coded, building a block ofinformation + parity (check) bits. All the bits are then codedwith a convolution code forming the coded bits.

8/13/2019 CSLGSM

46/87


Channel coding contd..

The two steps apply to both speech and data, though the

coding schemes are a little different. The reason for this double coding is that we want to

correct errors if we can (convolution coding) and afterthis we can detect (block coding) whether the informationis too damaged to use. If it is, it will be ignored.

The speech is divided into 20 ms segments. These 20ms speech pieces are digitized and then speech coded.The speech coder delivers 260 bits for each 20 ms ofspeech, which are divided into:

50 very important bits

132 important bits and 78 not so important bits.

8/13/2019 CSLGSM

47/87


Channel coding contd.. To the 50 bits, three parity bits are added (block

coding). These 53 bits together with the 132

important bits and 4 tail bits are convolution encodedto 378 bits (rate 1:2). The remaining bits are notprotected. This is illustrated in Figure below.

Figure 5: Channel coding in GSM.

8/13/2019 CSLGSM

48/87


Interleaving:

In real life, bit errors often occur in bursts. This is due tothe fact that long fading dips affect several consecutivebits.

Unfortunately, channel coding is most effective in

detecting and correcting single errors and error burststhat are not too long.

To deal with this problem we want to find a way ofseparating consecutive bits of a message, so that theseare sent in a non-consecutive way. This is done by

interleaving which principally means spreading outinformation.

An example of the method is given on the next slide.

8/13/2019 CSLGSM

49/87


Interleaving:Let us say a message block consists of four bits. We

take the first bit from four consecutive blocks and putthese numberone bits in a new block of four, whichwe can call a frame. The same thing is done with bits2 to 4 from the four message blocks. Then wetransmit the frames of numberone bits, numbertwobits etc.

Figure 6: Interleaving.

8/13/2019 CSLGSM

50/87


Interleaving:

During the transmission, frame 2 gets lost.Without interleaving, a whole message blockwould be lost, but in this case the interleavinghas the result that only the second bit in each

message block is faulty.

Figure 7: Received de-interleaved message blocks.

8/13/2019 CSLGSM

51/87


Interleaving:With channel coding, the information in all blocks canthen be restored. In GSM, the channel coder provides

456 bits for every 20 ms of speech. These areinterleaved, forming eight blocks of 57 bits each, seeFigure below.

Figure 8: Interleaving of 20 ms of encoded speech.

8/13/2019 CSLGSM

52/87

I t l i

8/13/2019 CSLGSM

53/87


Interleaving: If we take 2 x 57 bits from the

same speech frame and insertthem into the same burst, then

the loss of this burst (e.g. due tofading) would result in a totalloss of 25% missing bits, whichis too much for the channelcoding to cope with. So we haveto add another level ofinterleaving in between twospeech frames.

This will introduce a small delayin the system. On the otherhand, we can now afford to looseone whole burst since the lossonly affects 12.5% of the bits

from each speech frame, andwith channel coding we cancorrect this.

Figure 11: Second level of interleaving.

8/13/2019 CSLGSM

54/87


Modulation: The modulation method used in GSM is Gaussian Minimum

Shift Keying, (GMSK). It is a digital modulation form, i.e. the information to be sent isdigital. It could be data or digitized speech.

The modulator could be looked upon as a phase modulator.The carrier changes phase depending on the information bits

sent into the modulator. GMSK includes the desirable feature of a constant envelope

modulation within a burst. To get smooth curve shapes whenchanging the phase, the base band signal is filtered with aGaussian pass band.

With GMSK we get narrower bandwidth compared to ordinaryMSK, but the price for this is less resistance against noise.

Frequency hopping :

8/13/2019 CSLGSM

55/87


Frequency hopping:

In case of Rayleigh fading, the fading pattern is

frequencydependent. This means that thefading dips will occur at different places fordifferent frequencies.

We can simply say that to benefit from thisphenomenon, we change carrier frequencybetween a number of frequencies during the calland if only one of them has a fading dip, we willloose only a fraction of the information. Withcomplex signal processing, we can restore

information again. This technique is calledFrequency Hopping.

F h i d

8/13/2019 CSLGSM

56/87


Frequency hopping contd..During TDMA frame N, C 1 is used and during TDMA frame N+1, C2 is used. See Figure 12. The call, will use the same time slot butchange frequency in a know pattern which is repeated cyclically.

Figure 12: Frequency hopping between two frequencies C1 and C2.

8/13/2019 CSLGSM

57/87


Antenna (or space) diversity:

The risk of both of them being

affected by a deep fading dip at thesame time is small. This idea leadsto two Rx-antennas at the basestation, independently receiving thesame signal and thereforedifferently affected by fadingenvelopes.

By choosing the best of the twosignals, the degree of fading can bereduced on the uplink. At 900 MHz,it is possible to gain about 3 dB witha distance of 5 to 6 meters betweenthe antennas. At 1800 MHz due to

decreased wavelength the distancecan be shortened.Figure 13: Antenna diversity

8/13/2019 CSLGSM

58/87


The Equalizer:

Reflections from objects far away from the Rxantenna,i.e. time dispersion, causes Inter Symbol Interference,ISI.

The bits are spread out in time at the receiver and

adjacent symbols interfere with each other. This causes problems when the receiver tries to

understand which information has been sent.

The viterbi equalizer creates a mathematical model ofthe transmission channel, which in this case is the air

interface, and calculates the most probable transmitteddata.

8/13/2019 CSLGSM

59/87


The Equalizer:

Figure 14: Viterbi Equalizer

8/13/2019 CSLGSM

60/87


The Equalizer:

The data is transmitted in bursts which are placed withintime slots of length approximately 0.577 ms.

In the middle of each burst a known bit pattern called thetraining sequence is transmitted (S in Figure 14).

By comparing the known sequence, in a correlator at the

receiver with the received bit pattern (S), the equalizercreates a channel model.

A probable transmitted bit sequence is then fed throughthe channel model and the output is compared to thereceived bit sequence.

Th E li

8/13/2019 CSLGSM

61/87


The Equalizer: By comparing these two bursts the viterbi equalizer selects a

more probable transmitted bit pattern which is again fed throughthe channel model.

The process is repeated until a good enough bit pattern isfound.

The GSM specification prescribes an equalizer that should beable to handle a reflected signal delayed up to four bit times,

corresponding to about 15 ms, or a path difference betweendirect and reflected signal of about 4.5 km.

Now, the reflected signal too is influenced by Rayleigh fading dueto near region (from the receivers point of view) reflections.

So as long as the reflected signal is not delayed more than 15ms, we can simply say that it gives us some extra energy whichenhances the quality of the received signal.

Timing Advance:

8/13/2019 CSLGSM

62/87


Timing Advance: If the Mobile Station is moving away

from the Base Station during a call, it

will have to send the burst inadvance in relation tosynchronization time.

This in order for the burst to arrive inthe given time slot at the BaseStation (see Figure 15).

The Base Station will thereforecontinuously send a value between0 and 63, telling the MS how manybit-times (3,7 ms) ahead ofsynchronization time it shouldtransmit the burst.

This is one of the parameters whichputs a limit to how large a cell canbe.

Figure 15: Traffic on TS5, if themobile is moving away from the

BTS it will have to start sending

already during TS4.

8/13/2019 CSLGSM

63/87


The Air-Interface:

8/13/2019 CSLGSM

64/87


The channel concept:

The radio interface is the general name of the

connection between the mobile (MS) and the

base transceiver station (BTS).

In GSM it utilizes the TDMA-concept with oneTDMA-frame per carrier frequency.

Each frame consists of eight time slots (TS).

The direction from BTS to MS is defined as the

downlink and the opposite direction as theuplink.

8/13/2019 CSLGSM

65/87

Th h l t

8/13/2019 CSLGSM

66/87


The channel concept:Logical Channels:

A great variety of information must be transmitted betweenthe BTS and the MS, e.g. user data and control signaling.

Depending on the kind of information transmitted, we refer todifferent logical channels, i.e. the different types of informationare transmitted on the physical channels in a certain order.

These logical channels are mapped on to the physical

channels. For example speech is sent on the logical channelTraffic channel which during the transmission is allocated acertain physical channel, say TS 6 on carrier 0.

The logical channels are divided into two groups; cont ro lchannels and traff ic channels.

Th h l t

8/13/2019 CSLGSM

67/87


The channel concept:

Figure 17: Logical channels

C t l Ch l

8/13/2019 CSLGSM

68/87


Control Channels: When the MS is switched on it will be searching

for an adequate RBS to register with and listento. This is done by scanning the whole frequencyband, or, optionally, use a list containing certainfrequencies allocated for this operator.

When the MS has found the strongest carrier ithas to find out whether this is a BCCH carrier.

A BCCH carrier is the frequency used to carrythe broadcast channels. There is one per cell

and it is called c 0.

Broadcast Channels BCH:

8/13/2019 CSLGSM

69/87


Broadcast Channels ,BCH: Frequency Correction channelFCCH:

FCCH serves two purposes; one is to make sure this is

the BCCH-carrier, the other is to allow the MS tosynchronize to the frequency. FCCH is transmitted onthe downlink, point-to-multipoint.

Synchronization channelSCH:

Next thing for the MS is to synchronize to the TDMA

frame structure within this particular cell, and also tomake sure that the chosen base station is a GSM basestation. Listening to the Synchronization channel, SCH,the MS receives the TDMA frame number and also theBase Station Identity Code, BSIC, of the chosen base

station. BSIC can only be decoded if the base stationbelongs to the GSM network. SCH is transmitted on thedownlink, point to multipoint.

Broadcast Channels BCH:

8/13/2019 CSLGSM

70/87


Broadcast Channels BCH:Broadcast Control channelBCCH:

The last information the MS must receive in order to start

roaming, waiting for calls to arrive or making calls, is somegeneral information concerning the cell.

This is broadcasted on the Broadcast Control channel, BCCH,and does among others include the Location Area Identity(LAI), maximum output power allowed in the cell and theBCCH-carriers for the neighboring cells, on which the MS will

perform measurements. BCCH is transmitted on the downlink, point-to-multipoint.

Now the MS is tuned to a base station and synchronized withthe frame structure in this cell.

The base stations are not synchronized to each other, so

every time the MS decides to camp on another cell, its FCCH,SCH and BCCH have to be read.

Common control channels CCCH:

8/13/2019 CSLGSM

71/87


Common control channels, CCCH: Paging channelPCH:

Within certain time intervals the MS will listen to the Paging channel,

PCH, to see if the network wants to get in contact with the MS. Thereason could be an incoming call or an incoming Short Message. Theinformation on PCH is a paging message, including the MSs identitynumber (IMSI) or a temporary number (TMSI). PCH is transmitted onthe downlink, point-to-point.

Random Access channelRACH:

If listening to the PCH, the MS will realize it is being paged. The MS

answers, requesting a signaling channel, on the Random Accesschannel, RACH. RACH can also be used if the MS wants to get incontact with the network, e.g. when setting up a mobile originated call.RACH is transmitted on the uplink, point-to-point.

Access Grant channelAGCH:

The network assigns a signaling channel (the Stand alone DedicatedControl channel, SDCCH). This assignment is performed on the

Access Grant channel, AGCH. AGCH is transmitted on the downlink,point-to-point.

Dedicated control channels DCCH :

8/13/2019 CSLGSM

72/87


Dedicated control channels, DCCH : Stand alone Dedicated Control channelSDCCH:

The call set up procedure is performed on the SDCCH as well as thetransmission of textual messages (Short Message and Cell Broadcast)in idle mode. SDCCH is transmitted on both up- and downlink, point-to-point. The MS is on the SDCCH informed about which physical channel(frequency and time slot) to use for traffic (TCH).

Slow Associated Control channelSACCH:

On the uplink MS sends averaged measurements on own base station(signal strength and quality) and neighboring base stations (signalstrength). On the downlink the MS receives system information, which

transmitting power and what timing advance to use. SACCH istransmitted on both up- and downlink, point-to-point.

Fast Associated Control channel - FACCH

Handover is performed on FACCH. FACCH works in stealing mode,meaning that one 20 ms segment of speech is exchanged for signalinginformation necessary for the handover. The subscriber will notrecognize this interruption in speech since the speech coder will repeatthe previous speech block.

8/13/2019 CSLGSM

73/87

8/13/2019 CSLGSM

74/87

Burst:

8/13/2019 CSLGSM

75/87


Burst:Normal burst:

This burst is used to carry information on TCH andon the control channels BCCH, PCH, AGCH,SDCCH, SACCH and FACCH.

Figure 18: Normal Burst

The encrypted bits are 57 bits of encrypted data or speechplus one bit stealing flag indicating whether the burst was

stolen for FACCH signaling or not.

Burst:

8/13/2019 CSLGSM

76/87


Burst: The training sequence is a known bit pattern used by the

equalizer to create a channel model.

The reason why the training sequence is placed in the middleis that the channel is constantly changing. By having it there,the chances are better that the channel is not too differentwhen it affects the training sequence compared to when theinformation bits were affected.

The Tail Bits (TB) are always (000). They help the equalizer,

which needs a known start and stop bit pattern, i.e. thealgorithm used in the equalizer needs a certain start/stoppoint.

Since the time slot is 0,577 ms long, it has room for 156,25bits, but the burst only contains 148 bits. The rest of the space8,25 bits, is empty, and called Guard Period. 8.25 bits

corresponds to about 30 ms. The GP allows the transmitter toramp up and ramp down within limits specified by the GSMrecommendations.

Burst:

8/13/2019 CSLGSM

77/87


Burst: Access burst:

This burst is used for Random Access and has a longer guardperiod to cater for burst transmission from a mobile which does notknow the timing advance at the first access or after handover to anew BTS. The mobile may be far away from the BTS which meansthat the initial burst will arrive late and since there is no timingadvance on the first burst, it must be shorter to prevent it fromoverlapping the burst in the following time slot.

Figure 19: Access Burst

Dummy burst:This burst is sent from BTS on carrier c 0 when nothing else istransmitted. It carries no information. The format is the same as for the

normal burst with the encrypted bits exchanged for mixed bits with a

certain bit pattern.

The relationship between burst and

8/13/2019 CSLGSM

78/87



TDMA frame:

Figure 20: Basic TDMA frame, time slot, and burst structures


8/13/2019 CSLGSM

79/87



TDMA frame:

Hyperframes:

The cryptographic mechanisms use the TDMA

frame number as one parameter, so the BTS

must number the frames in a cyclic pattern (wecannot go on numbering them to infinity).

The chosen number is 2 715 648 which

corresponds to 3 h 28 min 53 s 760 ms.

This structure is named hyperframe

8/13/2019 CSLGSM

80/87

Mapping of logical channels on

8/13/2019 CSLGSM

81/87


Mapping of logical channels on

physical channels:

Figure 24: Full-rate and half-rate traffic channels

Figure 23: Mapping of TCH

Measurements of signal strength

8/13/2019 CSLGSM

82/87



by mobile station:

Measurements are performed in both idle mode (whenMS is switched on and moving around, roaming) and inactive mode.

Idle mode:

Cell selection is made at power on of the mobile.

1. The mobile scans all radio frequencies in the GSMsystem and calculates average levels for each of them.The mobile tunes to the strongest carrier and finds out ifit is a BCCH-carrier. If so, the mobile reads BCCH-datato find out if the cell can be locked to (chosen PLMN,

barred cell, etc.). Otherwise the mobile tunes to thesecond strongest carrier etc.


8/13/2019 CSLGSM

83/87



by mobile station:

Idle mode:2. The mobile may optionally include a BCCH-carrier

memory of valid BCCH-carriers in the home PLMN. Inthat case it will only have to search these carriers. If thisends unsuccessfully, the mobile performs as in 1.

On BCCH, the mobile is informed which BCCH-carriers itis to monitor for cell re-selection purposes. A list of thesix strongest carriers is updated regularly by the mobileas a result of the measurements.


8/13/2019 CSLGSM

84/87



by mobile station:

Active mode: During a call, the mobile continuously reports (via SACCH) to

the system how strong the received signal strength is from theBTSs in its surroundings.

These measurements are used by the BSC to make fast

decisions of target cells when a handover is required. The measurements on neighboring cells during a call takeplace when the mobile is not doing anything else, i.e. betweentransmission and reception on the allocated time slot.

The signal strength of the serving cell is monitored during thereception of the TS allocated to the mobile.


8/13/2019 CSLGSM

85/87



by mobile station:

Active mode: On SACCH, the mobile is informed which BCCH-carriers are

to be monitored for handover purposes, and the signalstrength of these are measured one by one.

The working schedule is therefore:

Transmit - measure - receive - transmit - measure - receive,and so on.

A mean value of the measurements for each carrier is thenderived and reported to the BSC.

Now, to be sure that the measured values corresponds withthe proper BTS, the identity of the BTS must also bedetermined. The identity of a BTS is given in BSIC, sent onSCH on c 0 , TS 0.


8/13/2019 CSLGSM

86/87



by mobile station:

Active Mode: 1. MS receives and

measures signalstrength and BER onserving cell, TS 2.

2. MS transmits. 3. MS measures signal

strength for at least oneof the neighboringcells.

4. MS tries to readBSIC on SCH (TS 0)for one of theneighboring cells.

Figure 25: MS measurement principle


8/13/2019 CSLGSM

87/87


by mobile station:

The six neighboring cells with highest mean signalstrength value and valid BSICs are then reported viaSACCH to BSC.

Since the MS might not be synchronized with aneighboring cell for which it is trying to determine theidentity, the MS does not know when TS 0 on thatBCCH-carrier will occur. Therefore it has to measureover a time period of at least 8 TS to be sure that TS 0will occur during the measurement.

This is accomplished with an IDLE frame as shown inFigure 25, step 4.

CSLGSM

Documents