Wcdma for umts radio access for third genergation mobile communacations phần 6 doc

Downlink Effects While the mobile in Figure 8.24 receives interference, it will also cause interference inuplink to Operator 2’s base station.. The worst-case adjacent channel interferen

The operator’s coverage probability requirement for the 8 kbps, 64 kbps and 384 kbpsservices was set, respectively, to 95 %, 80 % and 50 %, or better The planning phase startedwith radio link budget estimation and site location selections In the next planning step thedominance areas for each cell were optimised In this context the dominance is related only

to the propagation conditions Antenna tilting, bearing and site locations can be tuned toachieve clear dominance areas for the cells Dominance area optimisation is crucial forinterference and soft handover area and soft handover probability control The improvedsoft/softer handover and interference performance is automatically seen in the improvednetwork capacity The plan consists of 19 three-sectored macro sites, and the average sitearea is 7.6 km2 In the city area, the uplink loading limitation was set to 75 %, corresponding

to a 6 dB noise rise In case the loading was exceeded, the necessary number of mobilestations was randomly set to outage (or moved to another carrier) from the highly loadedcells Table 8.15 shows the user distribution in the simulations and the other simulationparameters are listed in Table 8.16

In all three simulation cases the cell throughput in kbps and the coverage probability foreach service were of interest Furthermore, the soft handover probability and loading resultswere collected Tables 8.17 and 8.18 show the simulation results for cell throughput

Figure 8.18 The network scenario The area measures 12 12 km2

and is covered with 19 sites, eachwith three sectors

and coverage probabilities The maximum uplink loading was set to 75 % according to Table8.16 Note that in Table 8.17 in some cells the loading is lower than 75 %, and,correspondingly, the throughput is also lower than the achievable maximum value Thereason is that there was not enough offered traffic in the area to fully load the cells Theloading in cell 5 was 75 % Cell 5 is located in the lower right corner in Figure 8.18, andthere is no other cell close to cell 5 Therefore, that cell can collect more traffic than the othercells For example, cells 2 and 3 are in the middle of the area and there is not enough traffic

to fully load the cells

Table 8.18 shows that mobile station speed has an impact on both throughput andcoverage probability When mobile stations are moving at 50 km/h, fewer can be served, thethroughput is lower and the resulting loading is higher than when mobile stations are moving

at 3 km/h If the throughput values are normalised to correspond to the same loading value,the difference between the 3 km/h and 50 km/h cases is more than 20 % The better capacitywith the slower-moving mobile stations can be explained by the better Eb=N0performance.The fast power control is able to follow the fading signal and the required Eb=N0 target isreduced The lower target value reduces the overall interference level and more users can beserved in the network

Table 8.15 The user distributionService in kbps Users per service

Table 8.16 Parameters used in the simulator

Slow (log-normal) fading correlation between base stations 50 %

Comparing coverage probability, the faster-moving mobile stations experience betterquality than the slow-moving ones, because for the latter a headroom is needed in the mobiletransmission power to be able to maintain the fast power control – see Section 8.2.1 Theimpact of the speed can be seen, especially if the bit rates used are high, because for low bitrates the coverage is better due to a larger processing gain The coverage is tested in thisplanning tool by using a test mobile after the uplink iterations have converged It is assumedthat this test mobile does not affect the loading in the network.

This example case demonstrates the impact of the user profile, i.e the service used and themobile station speed, on network performance It is shown that the lower mobile stationspeed provides better capacity: the number of mobile stations served and the cell throughputare higher in the 3 km/h case than in the 50 km/h case Comparing coverage probability, theimpact of the mobile station speed is different The higher speed reduces the required fast

Table 8.17 The cell throughput, loading and soft handover (SHO) overhead UL¼ uplink,

DL¼ downlinkBasic loading: mobile speed 3 km/h, served users: 1805

——————————————————————————————————————————Cell ID Throughput UL (kbps) Throughput DL (kbps) UL loading SHO overhead

Basic loading: mobile speed 50 km/h, served users: 1777Cell ID Throughput UL (kbps) Throughput DL (kbps) UL loading SHO overhead

Basic loading: mobile speed 50 km/h and 3 km/h, served users: 1802

Cell ID Throughput UL (kbps) Throughput DL (kbps) UL loading SHO overhead

fading margin and thus the coverage probability is improved when the mobile station speed

is increased

8.3.4 Network Optimisation

Network optimisation is a process to improve the overall network quality as experienced bythe mobile subscribers and to ensure that network resources are used efficiently Optimisa-tion includes:

1 Performance measurements

2 Analysis of the measurement results

3 Updates in the network configuration and parameters

The optimisation process is shown in Figure 8.19

A clear picture of the current network performance is needed for the performanceoptimisation Typical measurement tools are shown in Figure 8.20 The measurements can

be obtained from the test mobile and from the radio network elements The WCDMA mobilecan provide relevant measurement data, e.g uplink transmission power, soft handover rateand probabilities, CPICH Ec=N0and downlink BLER Also, scanners can be used to providesome of the downlink measurements, like CPICH measurements for the neighbourlistoptimisation

Table 8.18 The coverage probability results

Test mobile speed:

Test mobile speed:

Test mobile speed:

The radio network can typically provide connection level and cell level measurements.Examples of the connection measurements include uplink BLER and downlink transmissionpower The connection level measurements both from the mobile and from the network areimportant to get the network running and provide the required quality for the end users Thecell level measurements become more important in the capacity optimisation phase The cell

Performance analysis

Networks tuning

Key Performance Indicators (KPI)

Update of

parameters, site

configurations etc

Performance measurements

Figure 8.19 Network optimisation process

Figure 8.20 Network performance measurements

level measurements may include total received power and total transmitted power, the sameparameters that are used by the radio resource management algorithms.

The measurement tools can provide lots of results In order to speed up the measurementanalysis it is beneficial to define those measurement results that are considered the mostimportant ones, Key Performance Indicators, KPIs Examples of KPIs are total base stationtransmission power, soft handover overhead, drop call rate and packet data delay Thecomparison of KPIs and desired target values indicates the problem areas in the networkwhere the network tuning can be focused

The network tuning can include updates of RRM parameters, e.g handover parameters,common channel powers or packet data parameters The tuning can also include changes ofantenna directions It may be possible to adjust the antenna tilts remotely without any sitevisits An example case is illustrated in Figure 8.21 If there is too much overlapping of the

adjacent cells, the other cell interference is high and the system capacity is low The effect ofother cell interference is represented with the parameter other cell to own cell interferenceratio, i, in the load equations of Section 8.2, see Equation (8.16) The importance of the othercell interference is illustrated in Figure 8.22: if the other cell interference can be decreased

Figure 8.21 Network tuning with antenna tilts

=∑N

j =1

ηDL uj

Other cell interference

(Eb/N0)jW/Rj

If i can be reduced from 1.3 to 0.65, the number of users N can be increased 57 %.

We assume a = 0.5

[(1−α)+i

Figure 8.22 Importance of other cell interference for WCDMA downlink capacity

by 50 %, the capacity can be increased by 57 % The large overlapping can be seen from thehigh number of users in soft handover between these cells.

With advanced Operations Support System (OSS) the network performance monitoringand optimisation can be automated OSS can point out the performance problems, proposecorrective actions and even make some tuning actions automatically

The network performance can be best observed when the network load is high With lowload some of the problems may not be visible Therefore, we need to consider artificial loadgeneration to emulate high loading in the network A high uplink load can be generated byincreasing the Eb=N0target of the outer loop power control In the normal operation the outerloop power control provides the required quality with minimum Eb=N0 If we increasemanually the Eb=N0 target, e.g 10 dB higher than the normal operation point, that uplinkconnection will cause 10 times more interference and converts 32 kbps connection into

320 kbps high bit rate connection from the interference point of view The effect of higher

Eb=N0can be seen in the uplink load equation of Equation (8.12) The same approach can beapplied in the downlink as well in Equation (8.16) Another load generation approach indownlink is to transmit dummy data in downlink with a few code channels, even if there are

no mobiles receiving that data That approach is called Orthogonal Channel Noise Source,OCNS

For more information on the radio network optimisation process please refer to [3],Chapter 8, and for advanced monitoring and network tuning see [3], Chapter 10

8.4 GSM Co-planning

Utilisation of existing base station sites is important in speeding up WCDMA deploymentand in sharing sites and transmission costs with the existing second generation system Thefeasibility of sharing sites depends on the relative coverage of the existing networkcompared to WCDMA In this section we compare the relative uplink coverage of existingGSM900 and GSM1800 full rate speech services and WCDMA speech and 64 kbps and

144 kbps data services Table 8.19 shows the assumptions made and the results of thecomparison of coverage The maximum path loss of the WCDMA 144 kbps here is 3 dBgreater than in Table 8.4 The difference comes because of a smaller interference margin, alower base station receiver noise figure, and no cable loss Note also that the soft handovergain is included in the fast fading margin in Table 8.19 and the mobile station power class ishere assumed to be 21 dBm

Table 8.19 shows that the maximum path loss of the 144 kbps data service is the same asfor speech service of GSM1800 Therefore, a 144 kbps WCDMA data service can beprovided when using GSM1800 sites, with the same coverage probability as GSM1800speech If GSM900 sites are used for WCDMA and 64 kbps full coverage is needed, a 3 dBcoverage improvement is needed in WCDMA Section 12.2.1 analyses the uplink coverage

of WCDMA and presents a number of solutions for improving WCDMA coverage to matchGSM site density The comparison in Table 8.19 assumes that GSM900 sites are planned ascoverage-limited In densely populated areas, however, GSM900 cells are typically smaller

to provide enough capacity, and WCDMA co-siting is feasible

The downlink coverage of WCDMA is discussed in Section 12.2.2 and is shown to bebetter than the uplink coverage Therefore, it is possible to provide full downlink coveragefor bit rates 144 to 384 kbps using GSM1800 sites

Any comparison of the coverage of WCDMA and GSM depends on the exact receiversensitivity values and on system parameters such as handover parameters and frequencyhopping The aim of this exercise is to compare the coverage of the GSM base stationsystems that have been deployed up to the present with WCDMA coverage in the initialdeployment phase during 2002 The sensitivity of the latest GSM base stations is better thanthe one assumed in Table 8.19.

Since the coverage of WCDMA typically is satisfactory when reusing GSM sites, GSMsite reuse is the preferred solution in practice Let us consider next the practical co-siting ofthe system Co-sited WCDMA and GSM systems can share the antenna when a dual band orwideband antenna is used The antenna needs to cover both the GSM band and UMTS band.GSM and WCDMA signals are combined with a diplexer to the common antenna feeder.The shared antenna solution is attractive from the site solution point of view but it limits theflexibility in optimising the antenna directions of GSM and WCDMA independently.Another co-siting solution is to use separate antennas for the two networks That solutiongives full flexibility in optimising the networks separately These two solutions are shown inFigure 8.23 The co-siting of GSM and WCDMA is taken into account in 3GPP performancerequirements and the interference between the systems can be avoided

Table 8.19 Typical maximum path losses with existing GSM and with WCDMA

Receiver sensitivity1 110 dBm 110 dBm 125 dBm 120 dBm 117 dBm

Base station antenna gain4 16.0 dBi 18.0 dBi 18.0 dBi 1 8.0 dBi 18.0 dBi

Mobile antenna gain6 0.0 dBi 0.0 dBi 0.0 dBi 2.0 dBi 2.0 dBi

2

The WCDMA interference margin corresponds to 37 % loading of the pole capacity: see Figure 8.3

An interference margin of 1.0 dB is reserved for GSM900 because the small amount of spectrum in

900 MHz does not allow large reuse factors

8.5 Inter-operator Interference

8.5.1 Introduction

In this section, the effect of adjacent channel interference between two operators on adjacentfrequencies is studied Adjacent channel interference needs to be considered, because it willaffect all wideband systems where large guard bands are not possible, and WCDMA is noexception If the adjacent frequencies are isolated in the frequency domain by large guardbands, spectrum is wasted due to the large system bandwidth Tight spectrum maskrequirements for a transmitter and high selectivity requirements for a receiver, in the mobilestation and in the base station, would guarantee low adjacent channel interference However,these requirements have a large impact, especially on the implementation of a smallWCDMA mobile station

Adjacent Channel Interference power Ratio (ACIR) is defined as the ratio of thetransmission power to the power measured after a receiver filter in the adjacent channel(s).Both the transmitted and the received power are measured with a filter that has a Root-Raised Cosine filter response with roll-off of 0.22 and a bandwidth equal to the chip rate[11] The adjacent channel interference is caused by transmitter non-idealities and imperfectreceiver filtering In both uplink and downlink, the adjacent channel performance is limited

by the performance of the mobile In the uplink the main source of adjacent channelinterference is the non-linear power amplifier in the mobile station, which introducesadjacent channel leakage power In the downlink the limiting factor for adjacent channelinterference is the receiver selectivity of the WCDMA terminal The requirements foradjacent channel performance are shown in Table 8.20

GSM base

station

Dual band antenna for GSM and UMTS band

WCDMA base station

GSM base station

UMTS band

WCDMA base station

GSM band

Diplexer

Figure 8.23 Co-siting of GSM and WCDMA

Such an interference scenario, where the adjacent channel interference could affectnetwork performance, is illustrated in Figure 8.24 Operator 1’s mobile is connected to afar-away base station and is at the same time located close to Operator 2’s base station on theadjacent frequency The mobile will receive interference from Operator 2’s base stationwhich may – in the worst case – block the reception of its own weak signal.

In the following sections the effect of the adjacent channel interference in this interferencescenario is analysed by worst-case calculations and by system simulations It will be shownthat the worst-case calculations give very bad results but also that the worst-case scenario isextremely unlikely to happen in real networks Therefore, simulations are also used to studythis interference scenario Finally, conclusions are drawn regarding adjacent channelinterference and implications for network planning are discussed

8.5.2 Uplink vs Downlink Effects

While the mobile in Figure 8.24 receives interference, it will also cause interference inuplink to Operator 2’s base station In this section we analyse the differences between uplinkand downlink in the worst-case scenario The worst-case adjacent channel interferenceoccurs when a mobile in uplink and a base station in downlink are transmitting on full power,and the mobile is located very close to a base station that is receiving on the adjacent carrier

Table 8.20 Requirements for adjacent channel performance [11]

Adjacent carrier (5 MHz separation) 33 dB both uplink and downlink

Second adjacent carrier (10 MHz separation) 43 dB in uplink, 40 dB in downlink

(estimated from in-band blocking)

Weak signal

Operator

1

Operator 2

Interference

frequency

5 MHz 5 MHz

Adjacent channel interference

Figure 8.24 Adjacent channel interference in downlink

A minimum coupling loss of 70 dB is assumed here The minimum coupling loss is defined

as the minimum path loss between mobile and base station antenna connectors The level ofthe adjacent channel interference is calculated in Table 8.21 and it is compared to thereceiver thermal noise level of Table 8.22, both in uplink and in downlink The worst-caseincrease in the receiver interference level is calculated in Table 8.23

The maximum desensitisation in downlink is 41 dB and in uplink 22 dB, which indicatesthat the downlink direction will be affected before the mobile is able to cause highinterference levels in uplink This is mainly because of higher base station power compared

to the mobile power It is also preferable to cause interference to one connection in downlinkthan to allow that mobile to interfere with all uplink connections of one cell In the followingsections we concentrate on the downlink analysis

8.5.3 Local Downlink Interference

The adjacent channel interference in downlink may cause dead zones around interfering basestations In this section we evaluate the sizes of these dead zones as a function of the

Table 8.21 Worst-case adjacent channel interference level

Minimum coupling loss between mobile and 70 dB 70 dB

interfering base station in Figure 8.24

Adjacent channel interference 43 dBm 70 dB  33 dB 21 dBm 70 dB  33 dB

Table 8.22 Receiver thermal noise level

Thermal noise level kTB 108 dBm 108 dBm

Receiver noise level 108 dBm þ 7 dB 108 dBm þ 4 dB

coverage of the own signal The coverage is defined as the received pilot power level Theassumptions in the calculations are shown in Table 8.24.

The dead zones are evaluated as follows

1 Assume received pilot power level from Operator 1’s base station

2 Calculate maximum received signal power level for the voice connection In this case it isequal to the pilot power level since the maximum transmission power for voice isassumed to be equal to the pilot power of 33 dBm

3 Calculate maximum tolerated interference level I0 on the same carrier based on therequired Ec=I0

4 Calculate maximum tolerated interference level on the adjacent carrier based on theadjacent channel attenuation

5 Calculate minimum required path loss to the interfering base station

6 Calculate minimum required distance to the interfering base station

An example calculation is shown below assuming pilot power coverage of90 dBm

1 Assume pilot power level of90 dBm

2 Maximum received power for voice connection90 dBm

3 Maximum tolerated interference level I0¼ 90 dBm þ 18 dB ¼ 72 dBm

4 Maximum tolerated interference level on the adjacent carrier 72 dBm þ 33 dB ¼

39 dBm

5 Minimum required path loss 43 dBm (39 dBm) ¼ 82 dB when Operator 2’s basestation transmits with 43 dBm The required path loss is reduced to 33 dBm(39 dBm) ¼ 72 dB when operator 2’s base station transmits only common channelswith 33 dBm

6 Minimum required distance d¼ 10^ ((82  37)/20) ¼ 178 m or d ¼ 10^ ((72  37)/20)¼ 56 m

Table 8.24 Assumptions for dead zone calculation for 12.2 kbps voice

Transmission power of Operator 2’s base station 33–43 dBm

Pilot power from Operator 1’s base station 33 dBm

Maximum allocated power per voice connection from 33 dBm

Operator 1’s base station

Required Ec=I0for voice connection 7 dB 10log 10(3.84 e

6/12.2 e 3)¼ 18 dBPath loss calculation to the interfering Operator 2’s base 37 dBþ 20log 10ðdÞstation with distance d [metres] in line-of-sight

The results of the calculations are plotted in Figure 8.25 The results show that the deadzones can occur only if the following conditions take place at the same time: own networkcoverage is weak, the mobile is located close to the interfering base station that is operating

on the adjacent frequency with maximum power, and UE performance is just meeting 3GPPselectivity requirements

8.5.4 Average Downlink Interference

Since the probability of the adjacent channel interference is low, we need to resort to systemsimulations to evaluate the effect on the average performance More transmission power isneeded because of adjacent channel interference which leads to a lower capacity Thesimulations show the reduction in average capacity when the same outage probability ismaintained, with and without adjacent channel interference The simulation results andassumptions are presented in [12] The worst-case scenario is shown in Figure 8.26 wherethe site distance is 1 km and the interfering sites are just between our own sites The bestcase is when the operators’ sites are co-located

The simulation results are shown in Table 8.25 The worst-case capacity loss is 2.0–3.5 %.These capacity loss figures can be reduced with the solutions shown in the following section

8.5.5 Path Loss Measurements

The adjacent channel interference is basically about power competition between operators.The interference problems hit the connection if the interfering signal is strong at the sametime as the own signal is weak We can calculate the maximum tolerable power difference

Figure 8.25 Dead zone sizes as a function of own network coverage

between own signal and the interfering signal in Figure 8.24 When the maximum powerdifference is known, we can go and measure the power differences between two operators’networks and find the locations where the interference could cause problems We show anexample for WCDMA voice service in downlink with the following assumptions:

The required Ec=I0 for WCDMA voice¼ Eb=N0 – processing gain¼ 18 dB fromTable 8.24

The maximum transmission power per WCDMA connection is assumed to be 33 dBm

WCDMA mobile selectivity is 33 dB

The base stations’ transmit power is 43 dBm

The maximum allowed signal power difference between two operators can be estimated asfollows:

¼ Ec=I0þ mobile selectivity  downlink power allocation

¼ 18 dB þ 33 dB  10 dB ðthe power for a connection is 10 dB below the base stationmax powerÞ

¼ 41 dB

When the frequency separation is 10 MHz, the allowed signal power difference increases

to 51 dB Relative signal power measurements from today’s network show that theprobability of a larger power difference than 41 dB is typically <1–2 % and larger than

Table 8.25 Capacity loss because of adjacent channel interference

——————————————————————————————————————————

Own sites

Other operator sites (worst case)

Figure 8.26 Worst-case simulation scenario

51 dB is practically non-existent This is the probability that counter-measures are neededagainst interference The measurement results are in line with the simulation results.

8.5.6 Solutions to Avoid Adjacent Channel Interference

This section presents a few network planning and radio resource management solutions thatmake sure that adjacent channel interference does not affect WCDMA network performance

If the operators using adjacent frequency bands co-locate their base stations, either in thesame sites or using the same masts, adjacent channel interference problems can be avoided,since the received power levels from both operators’ transmissions are then very similar.Since there are no large power differences, the adjacent channel attenuation of 33 dB isenough to prevent any adjacent channel interference problems

The nominal WCDMA carrier spacing is 5.0 MHz but can be adjusted with a 200 kHzraster according to the requirements of the adjacent channel interference By using a largercarrier spacing, the adjacent channel interference can be reduced If the operator has twocarriers in the same base station, the carrier spacing between them could be as small as4.0 MHz, because the adjacent channel interference problems are completely avoided if thetwo carriers use the same base station antennas In that case a larger carrier spacing can bereserved between operators, as shown in Figure 8.27

In addition to the network planning solutions, the radio resource management can also beeffectively utilised to avoid the problems from inter-operator interference The calculations

in the sections above suggest the following radio resource management solutions to avoidadjacent channel interference in addition to the network planning solutions:

make inter-frequency handover to another frequency to provide higher selectivity andmore protection against interference;

allocate more power per connection in downlink to overcome the effect of theinterference;

Low interference

reduce the downlink instantaneous packet data bit rate to provide more processing gain totolerate more interference;

reduce the downlink AMR voice bit rate to provide more processing gain

8.6 WCDMA Frequency Variants

in Figure 8.28

The practical performance of WCDMA1900 using existing second generation sites in the

US is evaluated with a simulation case study in Section 8.6.3

8.6.2 Differences Between Frequency Variants

The main differences in the RF requirements between the frequency variants are summarisedbelow WCDMA2100 refers to WCDMA in the UMTS core band

Table 8.26 WCDMA frequency variantsFrequency variant Uplink [MHz] Downlink [MHz] Countries

Band I / UMTS core band 1920–1980 2110–2170 Europe, Asia, some

Latin Americancountries like Brazil

Latin Americancountries like Brazil

Asian countries

New channel numbers are defined Also, additional channels with 100 kHz raster aredefined for Bands IV, V and VI to allow WCDMA to be located exactly in the centre ofthe 5 MHz deployment in Figure 8.29 UMTS core band uses 200 kHz channel raster.

Narrowband blocking and intermodulation requirements are specified for mobile andbase stations to cope with the interference from the narrowband systems The requiredinterference rejection is 30 dB from a GSM carrier 2.7 MHz from the WCDMA centrefrequency in Figure 8.30 The narrowband blocking requirements are defined for theBands II, III, IV, V, where other technologies exist on the same band

MHz

IMT-2000 Downlink

USA

New 3G bands

(under discussion)

New 3G band Downlink

1700 1750

New 3G band Uplink

GSM1800 Uplink

GSM1800 Downlink

Figure 8.28 Spectrum for third generation services in the USA

WCDMA 5 MHz Other operator’s

narrowband systems GSM/TDMA/IS-95

Other operator’s narrowband systems GSM/TDMA/IS-95

5 MHz

Figure 8.29 Isolated 5 MHz allocation for WCDMA

30 dB attenuation with WCDMA mobile selectivity

The mobile reference sensitivity requirement is relaxed by 2–3 dB from 117 dBm to

115/114 dBm to allow high enough Duplex attenuation between uplink and downlink

in Bands II, III and V The separation between uplink and downlink is only 20 MHz inthose bands

These new requirements make the WCDMA deployment possible in an isolated 5 MHzblock shown in Figure 8.29 The inter-system interference in the 1.9 GHz band is verysimilar to the multi-operator interference that was discussed in Section 8.5, and the samesolutions can be applied If the operator has a 10 MHz continuous block, the inter-operatorinterference can be completely avoided by allocating WCDMA in the middle of the 10 MHzblock and narrowband 200 kHz GSM/EDGE carriers on both sides of the WCDMA Thenarrowband GSM/EDGE carriers protect the WCDMA carrier from the inter-operator inter-ference This approach is referred to as a sandwich approach and is shown in Figure 8.31

8.6.3 WCDMA1900 in an Isolated 5 MHz Block

The performance of WCDMA in an isolated 5 MHz block is evaluated in this section Theevaluation is based on a simulation case study using existing cell sites in a US network Thestudy area is a suburban area with 16 sites, each with three sectors, totalling 48 sectors.The average site covered 7 km2 The other operator’s site locations are randomly selectedtypical site locations between our own sites The results are presented in more detail in [13].The effect of the inter-operator interference to the capacity is studied and compared to theresults in [14]

The main interference mechanism is the downlink adjacent channel interference from theinterfering base station transmission to the WCDMA mobile reception We assume here thatthe adjacent operator uses GSM technology and the GSM sites are transmitting at 43 dBmcontinuously with an average antenna height of 25 m The simulation results are shown inTable 8.27

10 MHz

12 GSM/EDGE carriers

in 2.5 MHz

WCDMA

5 MHz

12 GSM/EDGE carriers

in 2.5 MHz Figure 8.31 10 MHz sandwich for WCDMA and GSM/EDGE

Table 8.27 WCDMA1900 simulation in an isolated 5 MHz block

Results from [13], Realistic scenario Results from [14], Worst-case scenario

The capacity loss shown in [13] is negligible The results in 3GPP report [14] show highercapacity loss than the results in [13] The target in 3GPP simulations has been to study theworst-case interference scenario where all the interfering sites are located at the edge of theWCDMA cells In that case the capacity loss is 1–2 %.

Finally, note that the inter-system interference problems can be completely avoided whenco-siting with other operators or when using a sandwich approach

References

[1] Sipila¨, K., Laiho-Steffens, J., Ja¨sberg, M and Wacker, A., ‘Modelling the Impact of the Fast PowerControl on the WCDMA Uplink’, Proceedings of VTC’99, Houston, Texas, May 1999, pp 1266–1270

[2] Ojanpera¨, T and Prasad, R., Wideband CDMA for Third Generation Mobile Communications,Artech House, 1998

[3] Laiho, J., Wacker, A and Novosad, T., Radio Network Planning and Optimisation for UMTS, JohnWiley & Sons, 2001

[4] Saunders, S., Antennas and Propagation for Wireless Communication Systems, John Wiley & Sons,1999

[5] Wacker, A., Laiho-Steffens, J., Sipila¨, K and Heiska, K., ‘The Impact of the Base StationSectorisation on WCDMA Radio Network Performance’, Proceedings of VTC’99, Amsterdam,The Netherlands, September 1999, pp 2611–2615

[6] Sipila¨, K., Honkasalo, Z., Laiho-Steffens, J and Wacker, A., ‘Estimation of Capacity and RequiredTransmission Power of WCDMA Downlink Based on a Downlink Pole Equation’, Proceedings ofVTC2000, Spring 2000

[7] Wang, Y.-P and Ottosson, T., ‘Cell Search in W-CDMA’, IEEE J Select Areas Commun., Vol 18,

No 8, 2000, pp 1470–1482

[8] Lee, J and Miller, L., CDMA Systems Engineering Handbook, Artech House, 1998

[9] Wacker, A., Laiho-Steffens, J., Sipila¨, K and Ja¨sberg, M., ‘Static Simulator for Studying WCDMARadio Network Planning Issues’, Proceedings of VTC’99, Houston, Texas, May 1999, pp 2436–2440

[10] Nokia NetActTMPlanner, http://www.nokia.com/networks/services/netact/netact_planner/[11] 3GPP Technical Specification 25.101, UE Radio Transmission and Reception (FDD)

[12] 3GPP Technical Report 25.942, RF System Scenarios

[13] Holma, H and Velez, F ‘Performance of WCDMA1900 with 5-MHz Spectrum Reusing 2G Sites’,presented at VTC’02 Fall, Vancouver, Canada, 24–29 September 2002

[14] 3GPP Technical Report 25.885 ‘UMTS1800/1900 Work Item Technical Report’

Radio Resource Management

Harri Holma, Klaus Pedersen, Jussi Reunanen, Janne Laakso

and Oscar Salonaho

9.1 Interference-Based Radio Resource Management

Radio Resource Management (RRM) algorithms are responsible for efficient utilisation ofthe air interface resources RRM is needed to guarantee Quality of Service (QoS), tomaintain the planned coverage area, and to offer high capacity The family of RRMalgorithms can be divided into handover control, power control, admission control, loadcontrol, and packet scheduling functionalities Power control is needed to keep theinterference levels at minimum in the air interface and to provide the required quality ofservice WCDMA power control is described in Section 9.2 Handovers are needed incellular systems to handle the mobility of the UEs across cell boundaries Handoversare presented in Section 9.3 In third generation networks other RRM algorithms – likeadmission control, load control and packet scheduling – are required to guarantee the quality

of service and to maximise the system throughput with a mix of different bit rates, servicesand quality requirements Admission control is presented in Section 9.5 and load control inSection 9.6 WCDMA packet scheduling is described in Chapter 10

The RRM algorithms can be based on the amount of hardware in the network or on theinterference levels in the air interface Hard blocking is defined as the case where thehardware limits the capacity before the air interface gets overloaded Soft blocking is defined

as the case where the air interface load is estimated to be above the planned limit Thedifference between hard blocking and soft blocking is analysed in Section 8.2.5 It is shownthat soft blocking based RRM gives higher capacity than hard blocking based RRM If softblocking based RRM is applied, the air interface load needs to be measured Themeasurement of the air interface load is presented in Section 9.4 In IS-95 networks RRM

is typically based on the available channel elements (hard blocking), but that approach is notapplicable in the third generation WCDMA air interface, where various bit rates have to besupported simultaneously

Typical locations of the RRM algorithms in a WCDMA network are shown in Figure 9.1

WCDMA for UMTS, third edition Edited by Harri Holma and Antti Toskala

# 2004 John Wiley & Sons, Ltd ISBN: 0-470-87096-6

9.2 Power Control

Power control was introduced briefly in Section 3.5 In this chapter a few important aspects

of WCDMA power control are covered Some of these issues are not present in existingsecond generation systems, such as GSM and IS-95, but are new in third generation systemsand therefore require special attention In Section 9.2.1 fast power control is presented and inSection 9.2.2 outer loop power control is analysed Outer loop power control sets the targetfor fast power control so that the required quality is provided

In the following sections the need for fast power control and outer loop power control isshown using simulation results Two special aspects of fast power control are presented indetail in Section 9.2.1: the relationship between fast power control and diversity, and fastpower control in soft handover

9.2.1 Fast Power Control

In WCDMA, fast power control with 1.5 kHz frequency is supported in both uplink anddownlink In GSM, only slow (frequency approximately 2 Hz) power control is employed InIS-95, fast power control with 800 Hz frequency is supported only in the uplink

9.2.1.1 Gain of Fast Power Control

In this section, examples of the benefits of fast power control are presented The simulatedservice is 8 kbps with BLER¼ 1% and 10 ms interleaving Simulations are made with andwithout fast power control with a step size of 1 dB Slow power control assumes that theaverage power is kept at the desired level and that the slow power control would be able toideally compensate for the effect of path loss and shadowing, whereas fast power control cancompensate also for fast fading Two-branch receive diversity is assumed in the Node B ITUVehicular A is a five-tap channel with WCDMA resolution, and ITU Pedestrian A is a two-path channel where the second tap is very weak The required E =N with and without

Figure 9.1 Typical locations of RRM algorithms in a WCDMA network

fast power control are shown in Table 9.1 and the required average transmission powers inTable 9.2.

Fast power control gives clear gain, which can be seen from Tables 9.1 and 9.2 The gainfrom the fast power control is larger:

for low UE speeds than for high UE speeds;

in required Eb=N0 than in transmission powers;

for those cases where only a little multipath diversity is available, as in the ITUPedestrian A channel The relationship between fast power control and diversity isdiscussed in Section 9.2.1.2

In Tables 9.1 and 9.2 the negative gains at 50 km/h indicate that an ideal slow powercontrol would give better performance than the realistic fast power control The negativegains are due to inaccuracies in the SIR estimation, power control signalling errors, and thedelay in the power control loop

The gain from fast power control in Table 9.1 can be used to estimate the required fastfading margin in the link budget in Section 8.2.1 The fast fading margin is needed in the UEtransmission power for maintaining adequate closed loop fast power control The maximumcell range is obtained when the UE is transmitting with full constant power, i.e without thegain of fast power control Typical values for the fast fading margin for low mobile speedsare 2–5 dB

9.2.1.2 Power Control and Diversity

In this section the importance of diversity is analysed together with fast power control Atlow UE speed the fast power control can compensate for the fading of the channel and keep

Table 9.1 Required Eb=N0values with and without fast power control

Slow power control Fast 1.5 kHz power Gain from fast

Table 9.2 Required relative transmission powers with and without fast power control

Slow power control Fast 1.5 kHz power Gain from fast

the received power level fairly constant The main sources of errors in the received powersarise from inaccurate SIR estimation, signalling errors and delays in the power control loop.The compensation of the fading causes peaks in the transmission power The received powerand the transmitted power are shown as a function of time in Figures 9.2 and 9.3 with a UEspeed of 3 km/h These simulation results include realistic SIR estimation and power controlsignalling A power control step size of 1.0 dB is used In Figure 9.2 very little diversity isassumed, while in Figure 9.3 more diversity is assumed in the simulation Variations in thetransmitted power are higher in Figure 9.2 than in Figure 9.3 This is due to the difference inthe amount of diversity The diversity can be obtained with, for example, multipath diversity,receive antenna diversity, transmit antenna diversity or macro diversity.

With less diversity there are more variations in the transmitted power, but also the averagetransmitted power is higher Here we define power rise to be the ratio of the averagetransmission power in a fading channel to that in a non-fading channel when the receivedpower level is the same in both fading and non-fading channels with fast power control Thepower rise is depicted in Figure 9.4

The link level results for uplink power rise are presented in Table 9.3 The simulations areperformed at different UE speeds in a two-path ITU Pedestrian A channel with average