Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION
1
Contents
Radio Network Optimization
• Reasons for the Need of Optimization
• Performance Data Measurements
• Drive Tests
• OMC-R counters analysis
• Interfaces types
• Radio interface analysis
• Analysis process
• Optimization of Physical Parameters
• Frequency planning & interference analysis
• Optimising adjacencies & LAC Location Area
• Optimising parameters O&M based performance parameters
• Optimization of Database Parameters
• Radio Link Failure (RLF) and Radio Link Timeout (RLT)
• Handover and Power Control Level triggered handover
• Level triggered Power Control
• Power Control Execution
• Quality triggered handover
• Quality triggered Power Control
• Handover triggered by power budget
• BCCH allocation
• Decision Process
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 2
1. Reasons for the Need of Optimization
Network optimization is an iterative process which should improve the quality and
performance of a network and also run the network more efficiently. As in any
optimization problem, also in network optimization, the network will mostly not run
optimal from the very beginning.
There can be mentioned several reasons:
• Systematic inaccuracies
• Statistical nature of the involved processes like e.g. traffic and RF propagation
• Dynamical nature of the involved processes like e.g. change of the subscriber’s
telephone behaviour (e.g. SMS)
• Wrong (or only too rough) planning assumptions, input data and/or planning models
• Increasing number of subscribers
• Installation errors (for example a wrong cabling: transmitting into cell A, but
receiving from cell B)
• Hardware / software trouble
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 3
2. Performance Data Measurements Performance data measurements can help the network operator for example to
localise problem areas as early as possible and also to verify improvements of the
network optimisation.
Concerning radio network optimisation there are related performance data
measurements foreseen by GSM (see: GSM 12.04) and in addition also vendor
specific ones.
In general performance data measurements can be run continuously, periodically or
sporadically, for a long time or a short time, observing smaller or greater parts of the
network.
The related counters could in principle be actualised continuously during the
observation period, but mostly a scanning method is used. Scanning method means
that the system counts the number of events not continuously but only at particular
times. This leads to some uncertainty for the measurement results. Neverthe less, the
error performed can be estimated using statistical methods.
In general, the smaller the scanning interval the higher the precision of the
measurement (for constant observation periods). Typical scanning intervals are 100
ms or 500 ms.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 4
3. Drive Tests Drive tests are performed by the network operator for various reasons:
• To check the coverage in a certain area
• To check the quality of service in a certain area
• To find the answer for customer complaints
• To realise that the network is not properly running
• To verify that the network is properly running
• To verify that certain optimisation steps have been successful
• ...
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 5
Drive tests must be well prepared. Before, during and after the drive test the following
steps should be performed:
Before drive test
• Plan the route where to drive.
• Plan the time when to drive.
• Determine the MS mode (idle/connected mode).
and also the call strategy (long/short calls).
• Decide which values to focus on (for example
:RXQUAL, RXLEV, SQI…).
• Select an appropriate test equipment and check the
test equipment.
• Think of notes which should be inserted later on in
the recording file.
During drive test
• Monitor the test equipment.
• Reconnect dropped calls .
• Insert notes in the recording file
• Note interesting events separately (e.g: on a piece
of paper)
After drive test
• Make back-up files of the capture data.
• Replay the captured data and analyse them
• Find out problem areas and problem events
• Use further post-processing tools to display the
captured data more
• clearly and to graphically display further values
• Perform statistics and summarise the results
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 6
Ericsson TEMS tool example
A measurement chain using TEMS includes:
• PC software with a serial port for the data.
• TEMS mobile with trace including the related software.
• GPS receiver.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 7
TEMS collected information
• Serving cell and neighbouring cells identities (BSICs) and BCCH frequencies.
• Radio parameters: RXLev, RXQual, TXPower, DTX, Timing Advance, FER, SQI
(voice quality), C1, C2,
• Current channel: CGI (MCC, MNC, LAC, CI), BSIC, BCCH ARFCN, TCH ARFCN,
Time slot, Channel type, Channel mode (FR, EFR, HR), Hopping Channel,
Hopping Frequencies, HSN (Hopping Sequence Number).
• Map to display the measurement itinerary with: parameters values, main events
(handover, call drop) and sites position. A GPS receiver is required for this
feature. • Level 2 messages (RR-RSP, DISC-CMD, UA-RSP, SABM-CMD...) and 3 (Synch
Channel Information, System Information Type 6, Measurement Report, Synch
Channel Information, Paging Request, Assignment Complete, Handover
Complete, ). • Frequency scanning.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 8
3. OMC-R counters analysis
Counters transmitted by the BSCs to the OMC-R.
Essentials to analyse the quality, to detect problems, to realize statistics, … at the
system side.
Analysis tools use these counters (generally, these are specific).
Example: Alcatel RNO or NPA, Metrica.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 9
4. Interfaces types
Air Interface (Um)
Provide information on the downlink as well as on the exchanged messages during
the protocols operation (calls, lcoation updates, …).
Tools (mobiles with trace and associated tools) such as Ericsson TEMS.
BTS-BSC Interface (Abis)
Allows evaluate radio performance of one or several calls in both ways (uplink and
downling). Allows observe resource allocation mechanisms (TCH or SDCCH) as well
as intra-BSC handovers operation.
Tools (protocol analysers) such as Siemens K11XX or K12XX series.
BSC-MSC Interface (A)
Allows capture additional information on the protocol operation and BSS - NSS
problems.
Tools such as Siemens K11XX or K12XX series.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 10
5. Radio interface analysis • Radio interface analysis tools: essential to identify the origins of the problems
(handover failures, coverage holes, bad quality due to interference, call drop,
…).
• Mobiles with trace: display the serving cell frequency, the allocated time slot
number, RXLEV and RXQUAL, neighbouring cells list, neighbouring cells
BCCHs, timing advance, ...
• Data can be stored in a laptop.
GPS receiver connection allows to display on a map (for instance in MAPINFO) the
mobile trajectory and the evolutions of the indicated parameters.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 11
Wanted
Information
Interface Protocol
Parameter
Message type
Identity of the BTS A, Abis Cell Identity (CI) in CM_SERV_REQ, PAG_RSP,LOC_UPD_REQ
Subscriber identity A, Abis IMSI/TMSI in CM_SERV_REQ, PAG_RSP,LOC_UPD_REQ
Actual LAC during LU A only LAI parameter in the Complete Layer 3 Information message (CLI3)
Former LAC during LU A, Abis LAI parameter i the LOC_UPD_REQ message
Sender/receiver of a SCCP message SCCP Called/Calling Party Address parameter (Cd/CgPA) in
the header of a SCCP message.
The Cd/CgPA consists of a combination of a point code, a subsystem number (HLR,VLR,MSC,etc) and a global title.
Are there any SCCP problems ? SCCP Look for CREF messages and UDTS messages.
If either message can be found, problems are certain
(overload?).
Check also if all CRs (Connection Request) are answered with CC (Connection Confirm)
Are there any SS7problems? SS7 Look for LSSU’s and COO’s (change over orders).
If LSSU’s (SIPO or SIB) are detected , then severe SS7 problems on one of the two ends of the SS7 link exist.
Are there any SS7problems because of high bit error rates ?
SS7, OMC Check if there have been frequent link failures recently.
If so, find out if the cause SUERM threshold exceeded is indicated.
Look for LSSU’s (SIO and SIOS) in the trace file.
Are there any problems in the VLR/HLR ? A, Abis Look for LOC_UPD_REJ, CM_SERV_REJ and AUTH_REJ messages.
Suspicious causes are: IMSI unknown in HLR, IMSI unknown in VLR, and LAC not allowed.
If this occurs frequently, then data errors in the NSS database are likely.
Is there any MS activity in a BSC or a BTS? A, Abis Look for CM_SERV_REQ, PAG_RSP and LOC_D_REQ.
Detection of CHAN_RQDo IMM_ASS_CMD is not sufficient.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 12
Are there any BSC problems? A Though related BTS’s do not suffer overload, there are many
ASS_FAI messages
Cause ‘33’=Radio Ressource Unavailable.
Is a BTS blocked ? Abis Check the RACH control parameters in the SYS_INFOS BCCH_INFOS 1-1. Is the cell Barr Access bit =1 or the Access Control Class not equal 0 ?
MSISDN/IMSI combinationof a subscriber MAP The BEG/provideRoamingNumber message possibly
contains both parameters.
Another possibility is the BEG/SendRoutingInformation message contains the MSISDN and the END/sendRoutingInformation message contains the IMSI
IMSI/TMSI combination of a subscriber A only Paging message (works on the A-Interface only)
Signaling Point Codes SS7 Routing Label in every message signal unit (MSU)
Distance between MS and BTS Abis Access delay in CHAN_RQD, timing advance (TA) in CHAN_ACT and all MES_RES.
For a conversion from TA to distance refer to the Glossary.
Target cell during handover A Cell Identity in HND_RQD messages
MS power class (Handly..) A, Abis Mobile Station Classmark X (RF Power Capability) parameter in CM_SERV_REQ, PAG_RSP, LOC_UPD_REQ
Called directory number in case of a MOC A, Abis Parameter Called Party BCD Number in SETUP message
Is DTX active ? Abis DTX(uplink): downlink measurements (MEAS_REP)
DTX (downlink): uplink measurements (MEAS_RES)
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 13
Are there Layer 1 problems on the Air-Interface?
Abis Look for CONN_FAIL messages (cause: ‘1’=Radio Interface
failure).
If this occurs frequently, then further investigation is necessary (e.g, identify affected TRX).
Are there Layer 2 problems on the Air-Interface?
Abis Look for ERR_IND messages (frequent cause: ‘1’= timer T200 expired (N200+1) times; ‘Chex’=frame not implemented)
Is there interference in the uplink or downlink ? Abis The RX_QUAL values are poor despite good or acceptable RX_LEV values in the uplink or downlink, frequent intra-BTS handover.
Check assignments rate.
Are there problems when sending TRAU frames between transcoder, BTS and MS
A, Abis Abis-Interface: Look for CONN_FAIL messages (cause:
‘28hex’ Remote Transcoder Alarm).
A-interface: Look for CLR_REQ messages (cause:
‘20’=Equipment Failure)
Are there problems during incoming Handovers?
A, Abis Abis-Interface: Look for CONN_FAIL messages (cause:
‘2’=Handover Access Failure).
A-Interface: Look for CLR_REQ messages ( cause: ‘0’=Radio Interface Failure)
Are there problems during outgoing Handovers?
A, Abis A-Interface: Look for HNBD_FAIL messages.
Abis-Interface: Look for HND_FAI messages.
Errors in the neighbourhood relations?
Poor overage?
A, Abis Check if there is hardly any outgoing handover. Check if the
number of CLR_REQ cause:
‘1’=Radio Interface Failure(A) and CONN_FAIL (cause: ‘1’=Radio Link Failure (Abis)) is higher than normal (location dependent).
Are there problems related to interworking between MSC and BSC ?
A High ASS_FAI rate. Causes : Requested Terrestrial Resources Unavailable, Terrestrial Circuit already allocated, Protocol Error BSC/MSC.
Check trunk assignment and other settings in MSC and BSC.
Were the BLO messages, possibly after a reset procedure, not repeated?
Are there any PLMN interworking problems ? MAP Many ABT messages from the affected PLMN (cause : Application Context Name not supported).
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 14
6. Analysis process
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 15
Benchmarking
• Troubleshooting : statistics, customer complaints
• Base station start-up
• Testing equipment :
Testing software e.g. TEMS
Test mobile phones (one or more)
Indoor/outdoor antenna
Cables + battery chargers
Field measurements • Testing route :
Roads, train
Hot spot, pedestrian
Urban, suburban, rural
• Test setup
Idle mode
Continuous call
Call sequence (90s calls / 15s idle)
tested frequencies: 900/900E/1800
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 16
Background On what basis does the customer rate GSM operators quality ?
• Prices
• Coverage area
• Call blocking/dropping
• Speech quality
• Customer service
• Else?
The user experienced service quality in GSM links directly to the performance of the
radio network Differentiation from competitors.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 17
7. Optimization of Physical Parameters
Altering antenna tilt
• to reduce interference.
• to limit coverage area.
• to improve coverage (e.g. coverage weakness below main lobe).
• to improve in-building penetration.
Altering the Antenna tilt must be done very carefully to really improve the situation.
Typical down-tilts are between 0° and 10°, however even higher values (up to 25°)
have already been used.
Altering antenna azimuth
• to overcome coverage weakness between different sectors.
• to reduce interference in certain directions.
Increasing or decreasing antenna height:
• to reduce or improve coverage.
• to reduce interference.
Change of antenna type
• to achieve desired ration characteristics.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 18
8. Interference reduction
Interference reduction means:
• Optimising is done frequency by frequency
• Antenna redirection
-the most predictable.
-coverage is in danger.
• Antenna tilting
- ideally very useful: steeping the slope.
- in practice difficult to predict.
should be always measured.
- useful for large clearance angles.
• Power reduction
- UL/DL interference power balance is lost.
- coverage is lost.
- not recommended.
All interference reduction means may generate problems in other frequencies.
Reducing interference Improves network quality !
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 19
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 20
9. Addition / re-movement of TRXs • Depending on the real measured traffic load either TRXs can be removed (switched
off or blocked) or must be added. Not really needed TRXs may interfere other cells.
• The number of needed TRXs and also the configuration of the different channels
depend on the offered traffic, and the subscriber behaviour.
10. Cell sectorization / cell splitting
Can be used for:
• Coverage enhancements (since the antenna gain of sectorised antennas is higher
than that of omni directional antennas)
• Interference reduction
• Capacity enhancements, but only if together with the sectorization also the number
of TRXs is increased (compare Erlang-B loss formula).
Depending on how the splitting is performed:
• It may be a more or less expensive and difficult (time consuming) solution.
• Coverage weakness between the main lobes may appear.
• The capacity will be reduced if the total number of TRXs remains constant.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 21
11. Implementation of Antenna near Pre-Amplifiers
Link imbalances are one reason for poor quality, increased call drop rate and
increased handover failure rate. In case of an unbalanced link, the uplink and
downlink coverage ranges differ. Often the downlink range is bigger than the uplink
range. This problem can be overcome by using antenna near preamplifiers which
improve the sensitivity and the noise figure of a base station system. Looking to the
link budget: The better the sensitivity of the base station, the more fare the possible
uplink range. In any case, a proper running network requires a balanced link.
Implementation of Repeaters
A repeater (see GSM 11.26 and GSM 05.05) is a bi-directional (full duplex) RF
amplifier and is used to overcome coverage holes in a base station area. Typical
applications of repeaters are the coverage of problem zones like tunnels, valleys, in
buildings, ...
A repeater receives, amplifies and retransmits the downlink signal from a donor base
station into an area with weak or no coverage, and the uplink signal from mobile
stations which are located in such an area. Repeaters extend but do not replace
base stations.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 22
12. Network quality benchmarking How does your networks quality match against the competitors ? and against the
world standard ?
Quality Urban Example benchmarking statistics
Call Success rate
Handover success rate
DL signal quality
DL signal power
MS TX power class_900
MS TX power class_1800
Handovers/ call SQI
95,70 % 97,90 % 0,4 -67.1 dBm 9,6 5,0 3,6 28,1
competitor 96,80 % 99,00 % 0,4 -61.6 dBm 5,5 3,5 3,5 28,8
Operator 1 97,90 % 99,20 % 0,2 - 8,7 1,8 2,5 29,1
Operator 2 96,90 % 95,50 % 0,5 - 7,2 1,5 1,4 28,2
Operator 3 95,40 % 96,10 % 0,6 - 7,8 N/A 3,1 27,3
Operator 4 95,20 % 95,00 % 0,3 - 8,7 N/A 1,7 28,9
Operator 5 95,20 % 99,10 % 0,5 - 7,3 1,4 2,4 28,8
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 23
Example:
Call Success Rate
SQI
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 24
13. Frequency planning & interference analysis
Frequency planning objectives
• QoS objectives
• Overall C/Ic and C/Ia requirements
• Cell-to-cell C/I requirements
Interference analysis
• Used by turns with the allocation in order to validate (and optimise, if
necessary) the frequency (reuse) plan.
Interference check per channel
• Evaluation of the significance of the residual interference
• HO/Cell selection margin should be included in the analysis
• C/Ic (and C/Ia ) statistics per channel
Interference per network
• On the dominance area
• C/Ic (and C/Ia ) statistics
• Shows the overall interference picture
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 25
Frequency Changes
• To overcome e.g. sever cases of downlink interference (therefore it is
advisable to have some spare frequencies).
• May influence other areas.
• Re-planning may become necessary.
• In high-density areas often difficult.
Strategies
• Using spare frequencies in severely interfered regions.
• TCH – BCCH change as temporary solutions in low TCH traffic load areas.
• Re-planning of TCH and BCCH frequencies.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 26
14. Optimising adjacencies & LAC
• Missing neighbours very often result in unnecessary dropped calls and bad
quality.
• Correct neighbour relations can be determined by analysing measurement
results.
Unnecessary LAC updates can easily increase
the signalling load significantly
Increased signalling reduces the room for payload
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 27
15. Location Area Location area (LA) planning plays an important role in cellular networks because of
the trade-off caused by paging and registration signalling. The upper boundary for
the size of an LA is the service area of a Mobile services Switching Center (MSC). In
that extreme case, the cost of paging is at its maximum but no registration is needed.
On the other hand, if each cell is an LA, the paging cost is minimal but the cost of
registration is the largest. Between these extremes lie one or more partitions of the
MSC service area that minimize the total cost of paging and registration. The
operator seeks to determine the location areas in an optimum fashion.
Too small ⇒ perhaps too many location updates.
(AGCH overload) (MS has to perform location update if location area is changed)
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 28
Too big ⇒ perhaps paging overload
(PCH overload) (MS is paged in the whole location area)
The size of the location area must always be a compromise:
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 29
16. Optimising parameters
study theoretically the impact of the intended change
select the set of performance parameters and test cells
change the cell parameters under optimisation
measure with a test MS
new conclusions
happy with the results ?
apply the changes permanently
in the network/service area
+
-
-
+
measure the performance and compare to the ref.
performance
happy with the results ?
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 30
17. O&M based performance parameters
QOS as well as RF and capacity planning parameters assessed per cell, per BSC
and per network
QOS parameters
• Dropped call rate
• Dropped call rate due to radio
• Cumulative UL/DL quality statistics
• TCH BH blocking rate
• SDCCH BH blocking rate
• (Call setup success rate)
• (Handover success rate)
RF planning parameters
• Number of calls
• TCH RF loss rate
• SDCCH RF loss rate
• TCH mean holding time
• Handovers per call
• Cumulative UL/DL level statistics
• Idle channel UL interference
• Power balance
Capacity planning parameters
• TCH BH traffic (e.g. weekly and average of daily BH traffics)
• BH activity per subscriber (segment)
• Total TCH time (per subscriber)
• SDCCH weekly BH traffic
• BH paging load
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 31
18. Optimization of Database Parameters
BSS PARAMETERS
Cell selection/reselection parameters:
• Cell_Reselect_Offset: Favor the cells of a frequency band.
• Temporary_Offset: Avoid Ping-Pong cell reselection.
• Cell_Reselect_Hysteresis: Avoid the reselection of cells belonging to different LAs
and reduces the unsuccessful paging rate. Example: 6 dB.
Access related parameters
• Max_Number_Retransmission: Maximum number of retransmissions on the
access channel (example: 1, 2, 4, 7). Default value: 2.
• Number_of_Slots_Spread_Trans: Maximum number of slots between 2
successives retransmissions (3 to 12, 14, 16, 20, 25, 32, 50).
• RXLEV_Access_Min: Defined the cell area. A change of 3 dB corresponds to
21% of the cell radius and 46% of the cell coverage area.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 32
19. Frequency Hopping Cyclic or pseudo random hopping?
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 33
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 34
20. Radio Link Failure (RLF) and Radio Link Timeout (RLT) Calls which fail due to radio coverage problems or which suffer under unacceptable
voice or data quality (due to e.g. interference) which cannot be improved by power
control or handover are either released or re-established in a defined way.
The criterion for the detection of a radio link failure by the MS is the success rate of
decoding DLSACCH messages.
The criterion for the determination of a radio link failure by the BS is either the
success rate of decoding UL-SACCH messages or it is based on RXLEV / RXQUAL
measurements.
The MS checks the DL with the help of a radio link (failure) counter running in the
MS.
The BS checks the UL with the help of a radio link (failure) counter running in the BS.
The algorithm for the modification of the radio link failure counter S is the following:
This algorithm is only running after assignment of a dedicated channel (i.e. in
connected mode).
The starting value Radio_Link_Timeout for the MS counter is sent on the BCCH
system information type 3 or on the SACCH system information type 6 in the
information element ‘Cell Options’.
Starting value for the Radio Link Failure Counter: Radio_Link_Timeout
In case of successful decoding of SACCH messages: Snew=Sold+2
In case of non-successful decoding of SACCH messages: Snew=Sold-1
Value range for S: 0<=S<=0 Radio_Link_Timeout
Radio link failure is detected if: S=0
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 35
21. Handover and Power Control
Reasons for the optimisation of handover parameters:
• To reduce the number of call drops
• To reduce the number of handovers
• To maximise the time duration the MS spends in the best cell : Target cell
correctly selected
• To improve the speech quality: Link quality maintained during the HO phase.
To meet these constraints, we shall minimise:
• The number of HO attempts
• The HO failure probability or call dropping rate,
• Ping-pong effect,
• Handover duration (handover triggering ;target BS link successful establishment),
• Resource consumption.
Handover types: intra- / inter- cell, BTS, BSC, MSC handovers
Handover causes:
• (Bad) RXQUAL
• (Low) RXLEV
• (far) DISTANCE
• (Power Budget) PBGT
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 36
Abbreviation
Remarks
L_RXLEV_UL_H
RXLEV threshold on the uplink for handover process to
commence (outgoing HO)
L_RXQUAL_UL_H
RXQUAL threshold on the uplink for handover process to
commence (outgoing HO)
L_RXLEV_DL_H
RXLEV threshold on the downlink for handover process to
commence (outgoing HO)
L_RXQUAL_DL_H
RXQUAL threshold on the downlink for handover process
to commence (outgoing HO)
MS_RANGE_MAX
Threshold for the maximum allowed distance between MS
and current BTS (outgoing HO)
RXLEV_UL_IH
RXLEV threshold on the uplink for intracell (interference)
handover
RXLEV_DL_IH
RXLEV threshold on the downlink for intracell
(interference) handover
RXLEV_MIN(n)
Minimum RXLEV required for a MS to be allowed to be
handovered to neighbour cell (n) (incoming HO)
HO_MARGIN(n)
Threshold for power budget process
MS_TXPWR_MAX
Maximum transmission power a MS may use in the serving cell
MS_TXPWR_MAX(n)
Maximum transmission power a MS may use in the adjacent cell (n)
BS_TXPWR_MAX
Maximum transmission power the BTS may use
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 37
22. Level triggered handover
Depending on the measured and averaged RXQUAL_XX and RXLEV_XX values the
system (MS and / or BS) may increase or decrease the output power or may
handover the call.
Remarks to the corresponding handover threshold settings:
L_RXLEV_UL_H and L_RXLEV_DL_H should be set some dB (e.g. 5 dB) above the
effective (+ diversity gain, + preamplifier) receiver sensitivity limit:
Receiver sensitivity levels due to GSM 05.05:
DCS 1800 class 1 or class 2MS 100 dBm
DCS 1800 class 3 MS 102 dBm
GSM 900 small MS 102 dBm
Other GSM 900 MS 104 dBm
Normal BTS 104 dBm
Example settings:
L_RXLEV_DL_H = –95 dBm
L_RXLEV_UL_H = –102 dBm
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 38
Note: There should be a level hysteresis between the threshold RXLEV_MIN(n) for
incoming handover and the threshold L_RXLEV_XX_H for outgoing handover:
RXLEV_MIN > L_RXLEV_XX_H + 4....10 dB
The size of this hysteresis should be related to the standard deviation of the long
term fading (typically 4...10 dB) and should be large enough to avoid ping-pong
handovers and small enough to allow fast handovers.
Example setting:
RXLEV_MIN = -90 dB
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 39
23. Level triggered Power Control Depending on the measured and averaged RX_QUAL and RX_LEV values the
system (MS and / or BS) may increase or decrease the output power or may
handover the call.
For the power control and handover threshold settings the following considerations
should be taken into account:
L_RXLEV_UL_P
(Lower) RXLEV threshold on the uplink for power increase
U_RXLEV_UL_P
(Upper) RXLEV threshold on the uplink for power reduction
L_RXLEV_DL_P
(Lower) RXLEV threshold on the downlink for power increase
U_RXLEV_DL_P
(Upper) RXLEV threshold on the downlink for power reduction
To avoid consecutive power increase or decreases directly after each other the
difference between upper and lower power control thresholds should be large
enough (e.g. 10 dB).
To allow the system to perform power control before handover is executed, the lower
power control level thresholds should be about 10 dB above the lower handover level
thresholds.
Example settings:
L_RXLEV_DL_H = –95 dBm,
L_RXLEV_DL_P = -85 dBm
U_RXLEV_DL_P = -75 dBm
L_RXLEV_UL_H = –102 dBm,
L_RXLEV_UL_P = -92 dBm,
U_RXLEV_UL_P = -82 dBm
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 40
24. Power Control Execution For the power control execution parameter settings the following considerations
should be taken into account:.
Since typically a power increase command is more urgent than a power reduction
command, the power increase step size should be greater than the power reduction
step size.
The power increase and power reduction step sizes should be on the one hand small
enough to enable an accurate power control, on the other hand large enough to
reduce the number of necessary power control commands and therefore the
signalling load.
Example settings:
POW_INCR_STEP_SIZE = 4 dB
POW_RED_STEP_SIZE = 2 dB
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 41
25. Quality triggered handover Depending on the measured and averaged RXQUAL_XX and RXLEV_XX values the
system (MS and/or BS) may increase or decrease the output power or may handover
the call.
Remarks to the corresponding handover threshold settings:
L_RXQUAL_UL_H, L_RXQUAL_DL_H
RXLEV_UL_IH, RXLEV_DL_IH
In case of bad quality (RXQUAL_XX>L_RXQUAL_XX_H) and high signal strength
(RXLEV_XX > RXLEV_XX_IH) at the same time, there is a high probability of the
presence of co-channel interference, adjacent channel interference, inter-modulation
problems, intersystem interference.
A Temporary solution should be intra-cell handover.
But Intra-cell handover doesn’t help: if frequency hopping is switched on, or if there is
only 1 TRX in the serving cell and the interference is continuous and not bursty.
Examples settings:
L_RXQUAL_UL_H = 5, L_RXQUAL_DL_H = 5
RXLEV_UL_IH = -85 dBm , RXLEV_DL_IH = -78 dBm
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 42
26. Quality triggered Power Control Depending on the measured and averaged RXQUAL_XX and RXLEV_XX values the
system (MS and/or BS) may increase or decrease the output power or may handover
the call.
Power is increased if the received quality is bad:
RXQUAL_XX > L_RXQAUL_XX_P
Power can be decreased if the received quality is very good:
RXQUAL_XX < U_RXQAUL_XX_P
However, often it is more suitable to control the power decrease by the level criteria
and to set U_RXQAUL_XX_P = 0 or a small value, i.e. to ‘disable’ the power
decrease due to good quality.
To make ‘power up before handover’ possible, the following relation between power
control and handover thresholds should be taken into account:
U_RXQUAL_XX_P < L_RXQUAL_XX_P < L_RXQUAL_XX_H
Example settings:
U_RXQUAL_XX_P = 0 (or 1)
L_RXQUAL_XX_P = 4
L_RXQUAL_XX_H = 5
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 43
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 44
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 45
27. Handover triggered by power budget In an interference limited area (e.g. small cells in cities) most of the handovers should
be power budget handovers:
For this type of handover not the level, quality, or distance is the handover cause,
since all the corresponding thresholds are not exceeded in the serving cell, but a
neighbour cell offers a better service (a smaller path loss, see link budget).
Since the power budget handover looks for the serving cell with the smallest path
loss, this kind of handover will:
• Reduce interference
• Prolong MS battery time
The power budget is defined as the difference between the path loss in the serving
cell and the path loss in the neighbour cell:
PBGT(n) = (BS_TXPWR – RXLEV_DL) – ( BS_TXPWR_MAX(n) – RXLEV_DL_NCELL(n))
Assumption:
BS_TXPWR_MAX– BS_TXPWR_MAX(n)=MS_TXPWR_MAX–MS_TXPWR_MAX(n)
PBGT(n) = RXLEV_DL_NCELL(n) – RXLEV_DL – PWR_C_D + min[MS_TXPWR_MAX,P] –
min [MS_TXPWR_MAX(n),P]
Where:
PWR_C_D is defined as: BS_TXPWR_MAX – BS_TXPWR
If PBGT(n) > HO_MARGIN(n) the path loss in the serving cell is greater than the path
loss in the neighbour cell + HO_MARGIN so that the neighbour cell is considered as
the much better cell.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 46
Remarks to the setting of the Handover Margin
• The HO_MARGIN setting should be a compromise between ideal power budget
handover (which requires a small HO_MARGIN value) and a setting to reduce the
risk of ping-pong handovers (which requires a greater HO_MARGIN value).
• A small handover zone increases the risk of ping-pong handovers.
• Usually HO_MARGIN is set symmetrically.
• Asymmetrical HO_MARGIN can be used to influence the size of the handover
area and/or to move the handover area, i.e. to move the cell boundaries.
• Adjusting HO_MARGIN values can therefore also be used to adapt the cell area
to the traffic load r to avoid local interference.
• RXLEV_MIN(n) should be set to such a value that RXLEV_DL_NCELL(n) >
RXLEV_MIN(n) in those areas where PBGT(n) > HO_MARGIN(n) to really allow
the power budget handover as soon as the power budget condition is fulfilled.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 47
Remarks to the pre-processing (averaging) of the measurements needed
for power control and handover decisions:
In general:
• Many measurements should be averaged in case that reliable decisions are
necessary (better statistics).
• Only a few measurements should be averaged in case that fast decisions are
necessary.
For level / quality triggered handover / power control decisions:
• To allow the system to ‘power up before handover’ usually the averaging process
for the handover decisions should include more measurements than for power control
decisions.
• Usually for level triggered handover decisions more measurement values should be
averaged than for quality triggered handover decisions since quality handovers must
be executed quickly if sudden interference appears.
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 48
28. BCCH allocation Also neighbor cell list is target of optimization process:
Missing neighbor cell ⇒ perhaps call drop
Too many neighbors ⇒ bad statistics, unprecise measurement values, perhaps
wrong decisions
In practice: ≈ 6-8 neighbors
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 49
Street corner effect: e.g. 20 dB loss
Fast handover mechanism necessary:
• trigger: uplink measurement receive level below threshold
• short averaging period of measurements
• predefined target cell lists
• small handover margins
• short timer settings
• allow back handover
Fast handover:
Radio Network Optimization
MN 1790-TN-09 RADIO PLANNING AND OPTIMIZATION 50
30. Decision Process
• The GSM network optimisation process consists of complex technical
measurements and analysis.
• Every network is individually planned and thus has individual problems which
decrease the quality.
• Professionally performed optimisation will increase customer’s satisfaction.
Problem Detection Solutions
Coverage • Access failure rate
• Call drop rate
• Important RXLEV HO rate
• New sites
• Antennas (tilt, azimuths, aperture)
• TMA installation
Interference • Communication quality
• RXQUAL HO rate
• Call dropping
• Frequency change
• Power control tuning
• Antennas action
Capacity • Blocking rate
• HO failure rate
• TRX adjunction
• Cell load distribution
• HO thresholds and cell access parameters adjustment
Handover Ping Pong
• Bad quality
• Micro-communication interruption
• HO parameters adjustement
• BTSs power adjustment