GSM: Radio interface

 

GSM: Radiointerface

---

One of the main objectives ofGSM is roaming. Thus, to allow for interoperability between MNsstations and disparate networks of the radio interface must bestandardised. Spectrum efficiency depends on aspects of the radiointerface and transmission, such as system capacity or techniquesused to optimize SIR and frequency reuse. It thus, becomes clearthat the specification of the radio interface can influence thespectrum efficiency.

Frequencyallocation

Two frequency bands, of 25Mhz each, are allocated for the GSM system:

• 890-915 Mhz for the uplink (MN to BTS).
• 935-960 Mhz for the downlink (BTS to M).

However, for reasons related tothe military as well as the existence of past analog systes (thatuse part of the two frequency bands), not all the countries can usethe whole GSM frequency bands.

Mediumaccess

GSM employs a mix ofFrequency Division Multiple Access (FDMA) and Time DivisionMultiple Access (TDMA), combined with frequency hopping.

Using FDMA, a frequency isassigned to each user. So for large number of usersin a FDMA system, the larger the number of requiredfrequencies. The limited available radio spectrum and the fact thata user will not free its assigned frequency until he does not needit anymore, reasons about scalability problems in an FDMAsystem.

TDMA allows several users toshare the same channel. Each subscriber multiplexes theshared channel, scheduling their frame for transmission. UsuallyTDMA is used with an FDMA structure.

In GSM, a 25 Mhz frequencyband is divided, using a FDMA scheme, into 124 carrierfrequencies with a 200khz spacing. Normally a 25 Mhz frequency bandcan provide 125 carrier frequencies; however, thefirst carrier frequency is used as aguard-band between GSM and other services working onlower freq. band. Each carrier is time-divided using a TDMA scheme.This scheme splits a 200khz channel, into 8bursts. A burst is the unit of time in a TDMA system,and it lasts approximately 0.577ms. Thus a TDMA lasts4.615ms. Each burst is assigned to a singleuser.
 

Channelstructure

A channel maps tothe recurrence of one burst every frame. It is definedby its frequency and the position of its corresponding burst withina TDMA frame. In GSM there are two types of channels:

• traffic channels used  for speech anddata.
• control channels used for network management messages andchannel maintenance tasks.

       Traffic channels (TCH)

Full-rate trafficchannels (TCH/F) are defined using a group of 26 TDMAframes called a 26-Multiframe. The 26-Multiframelasts 120 ms. In this frame group traffic channels for the downlinkand uplink are separated by 3 bursts. That implies,the mobiles will not need to transmit and receive atthe same time which simplifies considerably the electronics of thesystem.

The frames that form the26-Multiframe structure have different functions:

• 24 frames are reserved to traffic.
• 1 frame is used for the Slow Associated Control Channel(SACCH).
• The last frame is unused. It allows the MN to perform otherfunctions, such as measuring the signal strength of neighboringcells.

Half-rate trafficchannels (TCH/H), which double the capacity ofthe system, are also grouped in a 26-Multiframe but the internalstructure is different.
 
 

   Control channels

According to theirfunctions,  4 different classes of controlchannels are defined:
           Broadcast channels (BCH)

The BCH channels are used, byBTS to provide the MN with synchronization information from thenetwork. 3 different types of BCHs can be distinguished:

• Broadcast Control Channel (BCCH): gives to the MN theparameters needed to identify and access the network.
• Synchronization Channel (SCH): gives the MN the trainingsymbol sequence to demodulate the information transmitted byBTS.
• Frequency-Correction Channel (FCCH): provides the MNwith the frequency reference of the system for the purposes ofsyncronisation.

           Common Control Channels (CCCH)

The CCCH channels help toestablish the calls from the mobile station or thenetwork. These are:

• Paging Channel (PCH): used to alert the MN of anincoming call.
• Random Access Channel (RACH): used by the MN to requestnetwork access.
• Access Grant Channel (AGCH): used, by the BTS, to informthe MN about the channel it should use. This channel is the answerof a BTS to a RACH request from the MN.

          Dedicated Control Channels (DCCH)

The DCCH channels are usedfor message exchange between several mobiles or amobile and the network. These are:

• Standalone Dedicated Control Channel (SDCCH):used to exchange signaling in the downlink and uplink.
• Slow Associated Control Channel (SACCH): used forchannel maintenance and control.

           Associated Control Channels (ACCH)

Fast AssociatedControl Channels (FACCH) replace all or part of a trafficchannel when urgent signaling must betransmitted. The FACCH channels carry the same signaling as SDCCHchannels.
 
 

  Burststructure

Four different types ofbursts can be distinguished in GSM:

• Frequency-correction, used on the FCCH. It hasthe same length as the normal one but a differentstructure.
• Synchronization burst  used onthe SCH. It has the same length as the normal one but a differentstructure.
• Random access used on the RACH and isshorter than the normal burst.
• Normal burst used to carry speech or datainformation. It lasts approximately 0.577 ms and has a length of156.25 bits. Its structure is presented below.

 Structure of the 26-Multiframe, the TDMAframe and the normal burst

The tail bits(T) are a group of 3 bits set to zero and placed at thebeginning and the end of a burst. They coverthe periods of ramping up and down of the mobile'spower.

The coded data bitscorresponds to two groups, of 57 bits each,containing signaling or user data.

The stealing flags (S)indicate, to the receiver, whether the data bits are data orsignaling traffic.

The trainingsequence has a length of 26 bits. Itsynchronizes the receiver, thus masking out multipath propagationeffects.

The guardperiod (GP), with a length of 8.25 bits, isused to avoid a possible overlap of two mobilesduring the ramping time.

 Frequencyhopping

Propagation effects and thus,multipath fading depend on the radio frequency. To eliminatesignificant differences in channel quality, slow frequencyhopping is introduced; it changes the frequency with everyTDMA frame (fast frequency hopping changes thefrequency many times per frame but it is not used in GSM). Thefrequency hopping also reduces the effects ofco-channel interference.

There are different types offrequency hopping algorithms. The algorithm selected is sentthrough BCCH. Frequency hoping is optional for a BTS but must besupported by the MN.
 

From bits toradio

   The following figure shows the steps involved to transform speechaudio to radio waves and vice versa.
 
 

From bitsto radio

If the source of informationis data (not speech), the speech coding is notperformed.
 
 

   Speech coding

Talkspurts transmission(voice audio) is the mainstream service of a cellular system. TheGSM speech codec that transforms the analog signal (voice) into adigital representation, must meet the followingcriteria:

• Maintain speech quality at least equal to previous cellularsystems.
• Reduce redundancy in voice utterances. This reduction isessential due transmission capacity limitation on the datachannel.
• Adopt low complexity speech codec to reduce productioncosts.

The standard GSM speech codec isRPE-LTP (Regular Pulse Excitation Long-Term Prediction). This codecuses statistics from previous samples (information that doesn'tchange very quickly) to predict the current sample. The speechsignal is divided into blocks of 20ms. These blocks are thenpassed to the speech codec of 13 kbps,to obtain sppech frames of 260 bitseach.
 

  Channel coding

Channel coding addsredundancy bits to the original information to detect andcorrect, if possible,  transmissionerrors.

    Channel coding for the GSM data TCH channels

Channel coding is performedusing two codes: a block and a convolutionalcode.

The block codeis defined in the GSM Recommendations 05.03. It receives an inputblock of 240 bits and adds 4-zero tail bits at theend of the input block; this results a block output of244 bits.

A convolutionalcode adds redundancy bits to protect the data. Aconvolutional encoder contains memory. This propertydifferentiates the two types of code. A convolutional code can bedefined by three variables : n, k and K. Thevalue n corresponds to the number of output bits from theencoder, k to the number of input bits and K to thememory of the encoder. The ratio (R) of the code is definedas  R = k/n.

For example, a convolutionalcode with k=1, n=2 and K=5, uses a ratio ofR = 1/2 and delay of K=5, which meansthat it will add oneredundant bit for eachinput bit (1 in 2 output bits is an input bit). The codeuses 5 consecutive bits to compute the redundancybit. As the convolutional code is a 1/2 rate for aninput block of 244 bits an output blockof 488 bits is generated. These 488 bits arepunctured to produce a block of 456bits. 32 bits, obtained as follows, are nottransmitted :

                                                        C (11 + 15 j) for j = 0, 1, ..., 31

The output block of 456 bitsis then passed to the interleaver.
 
 

    Channel coding for the GSM speech channels

Before applyingchannel coding, the 260 bits of a GSM speech frame aredivided in 3 different classes according to function andimportance. The most important class is the class Iacontaining 50 bits. Next in importance is the classIb, which contains 132 bits. The least important isthe class II, which contains the remaining 78 bits.The different classes are coded differently:

• Class Ia bits areblock-coded. 3 parity bits are added to the 50class-Ia bits.
• The Ia output (53 bits) areadded to the Class Ib bits (50+3+132); 4 zero bits are addedto the Ia+Ib bits (185+4). A convolutional code, withr = 1/2 and K = 5, is then applied, obtaining an output block of378 bits (189*2).
• Class II bits are added(378+78), without any protection, to the output block of theconvolutional coder.
• The 456-bit block is finallyconstructed.

 

    Channel coding for the GSM control channels

In GSM, signallinginformation is contained in just 184 bits. 40 bitsparity, obtained using a fire code, and 4-zerobits are added to the 184 bits before applyingthe convolutional code (r = 1/2 and K = 5). The output of theconvolutional code is then a block of 456 bits; it does not need tobe punctured.
 

   Interleaving

This method rearranges agroup of bits in a particular way. It is combined with FEC codes inorder to improve the performance of the error correctionmechanisms. Interleaving decreases the possibility of losingwhole bursts during the transmission, bydispersing the errors. Since the errors becomeless concentrated, it is then easier to correctthem.
 

    Interleaving for the GSM control channels

At the physical layer a burstin GSM transmits 2 blocks of 57 data bits each. Thus, the 456-bitblock output of the channel coder fit into 4 bursts (4*114 =456). The 456 bits are, thus, divided into 8*57-bit blocks. Asinterleaving is applied during the forming of the blocks, the 1stblock of 57 bits contains the bit numbers (0, 8, 16, .....448), thesecond one the bit numbers (1, 9, 17, .....449), etc. The lastblock of 57 bits will then contain the bit numbers (7, 15,.....455).

The first 4 *57-bit blocksare placed in the even-numbered bits of four bursts.The other 4 are placed in the odd-numbered bits of thesame four bursts. Therefore, the interleavingdepth of GSM interleaving for control channels is 4and a new data block starts every 4 bursts.The interleaver for control channels is called a blockrectangular interleaver.
 

   Interleaving for the GSM speech channels

The 456-bit block, obtainedafter the channel coding, is divided in 8*57-bit blocks in the sameway as it is explained in the previous paragraph. But these 8blocks are distributed differently.

The first 4 blocks of 57 bitsare placed in the even-numbered bits of 4 consecutivebursts. The other four blocks are placed in the odd-numberedbits of the next four bursts. The interleaving depthof the GSM interleaving for speech channels is then 8. A newdata block also starts every 4 bursts. The interleaver for speechchannels is called a block-diagonalinterleaver.
 

   Interleaving for the GSM data TCH channels

A particular interleavingscheme, with an interleaving depth equal to 22, is appliedto the block of 456 bits obtained after the channel coding. Theblock is divided into 2 blocks of 6 bits each, 2 blocksof 12 bits each, 2 blocks of 18 bits each and 16blocks of 24 bits each. It is spread over 22 bursts in thefollowing way :

• the 1st and  22nd bursts carry one block of 6bits each (2)
• the 2nd and 21st bursts carry one block of 12 bits each(2)
• the 3rd and 20th bursts carry one block of 18 bitseach  (2)
• from 4th to 19th burst, a block of 24 bits is placed in eachburst (16)

A burst will then carryinformation from 5-6 consecutive data blocks. The datablocks are said to be interleaved diagonally. A new datablock starts every four bursts.
 

   Burstassembling

The burst assembling stepmanages the grouping the bits into bursts.
 

 Encryption

It is used to protectsignaling and data. An encryption key is computed using:

1. algorithm A8 (storedon the SIM card),
2. the subscriberkey
3. a random number (nonce)delivered by the network (same as the one used forauthentication).

A 114-bit sequence is producedusing:

1. the encryptionkey,
2. algorithmA5
3. the burstnumbers.

This bit sequence is thenXORed with the two 57 bit blocks of data included in anormal burst. To decrypt correctly, the receiver has to use thesame algorithm A5 for the deciphering procedure.
 

Modulation

The modulation chosen for theGSM system is the Gaussian Minimum Shift Keying(GMSK).

The GMSK modulation has beenchosen as a compromise between spectrum efficiency, complexity andlow spurious radiations (that reduce thepossibilities of adjacent channel interference). The GMSKmodulation has a rate of 270 5/6 kbauds and a BT product equal to0.3.
 
 

GMSKmodulator

Discontinuous transmission(DTX)

DTX is used to suspend theradio transmission during the silence periods. Thisexploits the observation that only 40-50% during a conversationdoes the speaker actually talk. DTX helps also to reduceinterference between different cells and to increase systemcapacity. It prolongs battery charge life. The DTX function isperformed by means of:

• Voice Activity Detection (VAD), which has todetermine whether the sound represents speech or noise, even if thebackground noise is very important. If the voice signal isconsidered as noise, the transmitter is turned offproducing then, an unpleasant effect calledclipping.
• Comfort noise.  A side-effect of theDTX function is that when the signal is considered as noise, thetransmitter is turned off and therefore, a total silence is heardat the receiver. This can be very annoying to the receiving usersince it appears  as a dead connection. In orderto overcome this problem, the receiver creates a minimum ofbackground noise called comfort noise. Comfort noiseeliminates the impression that the connection isdead.

Timingadvance

The timing of the burststransmissions is very important. Mobiles are at differentdistances from the BTS. Their delay depends, consequently,on their distance. Timing advance allows signals coming fromdifferent distances to arrive to the BTS at the right time. Thelatter measures the timing delay of the MNs. If the burstscorresponding to an MN arrive too late andoverlap with other bursts, the BTS tells, the MN to advancethe timing in transmission of its bursts.
 
 

Powercontrol

The BTSs perform timingmeasurements; they also perform measurements on the power level ofthe different mobile stations. These power levels are adjusted sothat the power is nearly the same for each burst.

The BTS controls its powerlevel. The MN measures the strength and the quality of the signalbetween itself and the BTS. If the mobile station does not receivecorrectly the signal, the BTS changes its power level andretransmits.
 
 

Discontinuousreception

It is a method used toconserve the MN's power. The paging channel is divided intosubchannels corresponding to single mobile stations. Each MN'listens' only to its subchannel while it stays insleep mode for the duration of the rest subchannels of thepaging channel.
 

Multipath andequalisation

At the GSM frequency bands,radio waves reflect from buildings, cars, hills, etc. So not onlythe 'right' signal (the output signal of the emitter) is receivedby an antenna, but also many reflected signals, which corrupt theinformation, with different phases.

An equaliser is in charge ofextracting the 'right' signal from the receivedsignal. It estimates the channel impulse response ofthe GSM system and then constructs an inverse filter.The receiver knows which training sequence it must wait for.By means of comparing the received training sequencewith the expected one, the receiver computes thecoefficients of the channel impulse response. Inorder to extract the 'right' signal, the received signal ispassed through the inverse filter.

  http://www.cs.ucl.ac.uk/staff/t.pagtzis/wireless/gsm/radio.html