Optimization can be defined as that activity which is performed to configure the existing radio network to deliver optimum performance. The goal of optimization is to provide the network performance to meet the targets set by the operator, with cost effective means. Network performance means here capacity and coverage provision in addition to the quality experienced by the end user.
Company confidential
Optimization
Company confidential
Data Collection Equipment req. Services test Call Patterns OSS Alarms Network Stats
Troubleshooting
Root Cause analysis
Why drive the network? New Site Integration RF Tuning Network Optimization Network Benchmarking
Company confidential
Survey
Drive survey
Company confidential
Cluster Acceptance
Types of measurements
Scanner system used in initial cluster optimization
To measure dominance, coverage, pilot pollution etc. Physical changes (antennae) in NW after data processing Logical changes (parameters) adjustments cell power setting neighbour list definition
Company confidential
UE Logging tools
TEMS Agilent 7464A XCAL-W TOM
Company confidential
UE
GPS
USB Hub
Company confidential
Take note of all the COM ports assigned for each device connected for UE and Scanner. Each connected device will be assigned an individual COM ports.
Company confidential
Company confidential
Set the Scanner Properties for the UMTS Channel, Number of Scrambling codes and the required Ec/No threshold.
Defined the scanner properties and start the scan to log the scan data.
Company confidential
Defining workspaces
Company confidential
Defining Workspaces
Create worksheet for CS for observing CPICH RSCP, Ec/Io, Radio Parameters, UE tx pwr. Create worksheet for signalling messages and events Create worksheet for displaying PS Data messages. Create worksheet for Scanner messages
Company confidential
Example of Events Worksheet containing Call Events, Layer 3 and Layer 2 Messages
Company confidential
Company confidential
Example of worksheet for a PS call containing Session Throughput displayed in a graphical form.
Company confidential
Company confidential
Create and load the cell definition file in the right format. This will display the site information and the change of the cells during the drive.
Defined the scanner properties and start the scan to log the scan data.
Company confidential
Create script for each test case scenario. Here is a sample script for voice call of duration 120s, wait time of 20s and run for 100 times. Similar scripts can be setup based on the test requirements for short call and long call.
Company confidential
Here is a sample script for downloading 2MB file from the FTP server TMO03.com for 10 times. Similar scripts can be setup based on the test requirements for other packet services.
Company confidential
Measurements should have depth and breadth. Measurement breadth is determined by the extents of the drive test campaign
Classification Downtown Suburban Rural Suggested Drive Breadth Every Street Every Block As required Comments In downtown areas it is possible that GPS may wander away from the streets
Measurement depth is determined by the number of measured servers at any particular location
Scanner measurements are preferred over UE measurements
Scanners are not constrained by a neighbor list Tend to have great equipment sensitivity
Company confidential
Company confidential
It is important that cells within the cluster are in normal working condition. Missing cells may have negative impact to the analysis process and optimization decisions Network availability is greater than 95%. Various OSS tools can be used to collect the network availability.
Company confidential
Company confidential
Alarms associated with BTSs should be checked before, during and after optimization (drive test) The rule of thumb is that the sites are free of any critical alarms before drivetest Alarms can relate to:
T1 errors TMA/Antenna/Feeder issues NodeB/RNC/Core network alarms External alarms
Alarm check can done with standard netact tools. It is engineers responsibilty to determine the impact of these alarms and ascertain if the drive test can proceed.
Company confidential
Parameter Check
Basic parameter checking such as CPICH power, Primary scrambling code and 2G / 3G neighbors should be performed to ensure these planned parameters match the RNC configuration.
In addition:
Barring / cell reservation parameters should be checked to ensure correct settings. OCNS parameters and site lists must be verified since loaded cluster drive tests are being performed. Uplink loading is being simulated by reducing the maximum allowed uplink transmit power by 3dB therefore this parameter needs to be checked too.
Company confidential
Data Collection
Once the above section is fulfilled without any errors, you can now proceed to collect drive test data. Always follow the agreed drive route unless there are unavoidable reasons that force you to change the route. If such a situation arises, log all these changes in the drive_data_sheet. Any changes must be properly communicated. Before starting the data collection, verify that the laptop system clock is set right and the right time zone. Load Predefined Workspace created during the installation and configuration process
Company confidential
Attach a script for each UE Observe that each sequence in the script is running properly Observe that signaling messages are captured from the UEs CPICH information is being captured by the scanner GPS is receiving satellite signal. GPS coordinates is registered Vehicle speed as defined in the drive route and as the law allows should always be observed. Avoid getting into wrong-way or construction areas. Details of any problem should be recorded in the drive test data sheet (e.g. no GPS signal, UE suddenly stopped working, loss of power, route not drivable, and etc)
Company confidential
Recording must be kept at minimum file size (eg:10MB) In special cases where the road has no outlet other than going back, a new file name should be used when going back In some places where traffic is so bad and vehicle is not moving, recording should be stopped and resume as soon the traffic moves Use specified file naming convention for TEMS logs as Date_Time_Market_Cluster_TestCase_Test#.dt1 Replay logs to verify its integrity at the end of each drive test.
Company confidential
Company confidential
Replay the log file and click on the events with failure indication for setup or handover. All the windows are in sync. So at a particular instant of time, check for details on RSCP, Ec/Io, UE Tx pwr, location of UE on the map, scanner information
Layer 3 message also gives the complete details of the information on what is happening in UL and DL.
Identify the cause, fix it and re-drive and confirm whether the failure is still occurring or not.
Company confidential
Layer 3 messaging Look for the details in the Events and Layer 3 messages
Company confidential
TEMS is having in-built report functionality which can provide quick snapshot on the performance of the drive test. Run the report from Log file/export generator. A summary will be listed on all Key indicators on success and failures for voice and packet.
Company confidential
Company confidential
Excercise
Setting up and saving custom workspaces in the drive test tool. Setup the drive test tool in the vehicle. Ensure all the equipments are connected and are providing valid measurements. Check for alarms and cell availabilty on the training network.
Perform scanner/UE based drive test on a pre-defined route in the training network.
Check data integrity and replay logfiles in TEMS.
Company confidential
Company confidential
Actix-Analyzer Classic
Call test data and scan data can be analyzed using map view and table view
Variety of reports can be run for call statistics, Handover Analysis of Signalling messages can also be done Lot of customization can be done on viewing the results
Company confidential
Company confidential
Create superstreams
Company confidential
Display the Scan plot for RSCP and Ec/Io to see the coverage and quality.
Company confidential
Display the Scan plot for RSCP to see the spread out for each scrambling code. By this the need for any down tilting can be identified for any overshooting sectors.
Company confidential
Analysis of Active Set RSCP_0 gives an idea on the strong and weak coverage areas for the cell in the drive route.
Company confidential
Analysis of Active Set Ec/Io gives an idea on the strong and weak Ec/Io (interference) areas for the cell in the drive route.
Company confidential
Analysis of events helps to focus on the areas where setup failures, dropped calls occur in the drive route and to dig further on the reasons by using other attributes.
Company confidential
Analysis of RSCP and Ec/Io helps to focus on the good coverage with bad qual and optimization can be focussed on this area.
From the view, forms can be opened which would be used during analysis.
Company confidential
Detailed Analysis on the signalling messages helps to understand the root cause of the problem and this also helps to understand the call setup process, measurement reporting, etc.
Company confidential
Lot of reports can be run using Analyzer to identify the problems on call setup, dropped call, HO performance
Company confidential
In the Analysis tab, the report for UMTS Drive Test Summary can be run. The distribution of of statistics Ec/Io, RSCP and call statistics can be obtained for each site.
Company confidential
In the result of the drive test summary, the report on Coverage Summary can be opened to see the distribution of Ec/Io and RSCP
Company confidential
In the Analysis tab, the report for Detailed call Setup analysis can be run to see the breakdown on the performance of call set-up.
Company confidential
In the Analysis tab, the report for Detailed call Setup Analysis can be run to see the breakdown on the performance of call set-up.
Company confidential
CPICH meausrement before RRC Connection Request- In case of failures, this report helps to identify the signal strength in the active and monitored sets before the failure.
Company confidential
Details of Call Setup Analysis; It is a very good snapshot to see the performance of the drive test
Company confidential
Troubleshooter is good option to investigate the performance details very quickly by selecting the required problem. The summary will be available on the dash board and this module will automatically pull the required plots from the drive data.
Company confidential
Troubleshooter simplifies the analysis and helps the Engineer to focus on critical analysis like call set up failures, important issues like poor coverage issues, pilot pollution and missing neighbours.
Company confidential
Troubleshooter provides the summary charts on KPIs for Call setup success, dropped call, Active set update, Handoff performance. The performance of each sector is also displayed.
Company confidential
If we select the Call setup failure investigation in the dashborad, this is the quick analysis on map display for Active set Ec/IO.
Events, neighbours, line chart on Ec/Io, signal strength, UE txpower and SIR are also displayed.
Company confidential
If we select the poor coverage analysis in the dashborad, this is the quick analysis on map display for good and poor rscp and ec/io areas, bar chart on active set and neighbours.
Company confidential
If we select the Handover Investigation analysis in the dashboaad, this is the quick analysis on map display for UE Handoff state whether it is single sector service or in soft HO state. Additional information is provided on neighbour list and line chart with key attributes.
Company confidential
If we select the Pilot Pollution investigation analysis in the dashboard, this is the quick analysis on map display for CPICH Scan Ec/Io, too many servers and pilot pollution events. Additional information is provided on top 10 Scanners.
Company confidential
Spotlight is also similar to Troubleshooter and this gives the following output. Summary Dashborad report on failure issues, coverage issues, pilot pollution.
Event explorer To focus specifically on call setup failures and drop calls.
Radio Network Explorer- To focus on the coverage, pilot pollution and missing neighbour analysis.
Company confidential
Options available from the dashboard of Spotlight for Event and Radio Network Explorer and summary on the issues.
Company confidential
Analysis on the performance on Dropped calls, Setup failures by selecting Event Explorer Map display with required attributes and events.
Company confidential
Analysis on coverage by selecting on the cell coverage from Radio Network Explorer Map display with required attributes and events.
Company confidential
Analysis on the performance on Cell pilot pollution by selecting Radio Network Explorer and cell pilot pollution.
Company confidential
ANALYSIS
Company confidential
Survey
Drive survey
Company confidential
Cluster Acceptance
RF Performance reporting
Plot of measured Best Server RSCP Plotted along the drive route of the map. Shows measured best servers CPICH RSCP level so that data analyst can see the coverage problem area.
Plot of measured best server Ec/No Plotted along the drive route of map. Shows measured best servers CPICH Ec/No level so that data analyst can see the dominance problem area (low best server Ec/No)
Histogram and CDF of RSCP. Shows the percentage of samples at different received RSCP level for whole drive data. Histogram and CDF of Ec/No
Shows the percentage of samples at different received Ec/No level for whole drive data.
Company confidential
Check that the RSCP levels for the area meet the target thresholds for the service and environment (clutter, building loss)
Company confidential
Overshooting Areas
Company confidential
Plotting the RSCP for individual scrambling codes can help in identifying hardware and data translation issues where the coverage area of the cell is different from the p-lanned. This can include swapped feeders, incorrect scrambling codes assigned to the cells, incorrect azimuths and tilts.
Company confidential
Ec/Io Verification
Company confidential
e.g criteria can be set as RSCP > -92 dbm and Ec/No < -9 dB
Company confidential
Look at each SC and determine the best way to optimize the area.
Company confidential
Company confidential
UL Coverage Verification
By using UE during testing we can check the UEs transmit power for problem areas where UE cannot maintain the link quality
Company confidential
A
1. 2. 3. 4. 5. Etc.
No
e.g. 98 dBm
Possible Actions/Solutions Antenna Tilting Antenna Panning Change Antenna Type Change Antenna Height Change CPICH Tx Pwr
Currently X= Max As size=3 for Nokia
A
e.g. 12 dB
No
Yes
Multi-path Problem A
No
In block A, check if the cause for call setup failure is CPICH RSCP and Ec/No coverage problem The received best servers CPICH RSCP and Ec/No will be compared to the coverage thresholds at the location where call setup failed and if best servers RSCP or Ec/No is less than the thresholds, coverage optimzation will be performed.
The proposed coverage improvement solution should be verified with simulations from a planning tool. This gives the approximation if the changes meet the expected results.
Company confidential
The area that does not meet one of the following criteria is defined as coverage problem area and dominance problem area respectively.
Company confidential
Tilt can be electrical or mechanical and both with their advantages or disadvantages. Tilting activity should be followed by a drive test to evaluate the impact of changes Decisions should be made based on scanner measurements. UE measurements can be used, but they can sometimes lead to wrong results like due to a missing neighbor.
Company confidential
Hard handovers are based upon either intra-frequency, inter-frequency or inter-system neighbour lists
IntraFrequency Neighbors
Cell a Cell b Cell c Cell d Cell e
InterSystem Neighbors
Cell r Cell s Cell t Cell u Cell v
The critical objective of neighbour list verification is to Ensure that sufficient neighbours are present to achieve the desired coverage performance
Max. 32
Each neighbour has a set of associated parameters e.g. CPICH measurement offset Actix Analyzer is able to suggest appropriate neighbour lists Strategy for initial system deployment is to place the emphasis upon adding neighbours rather than removing them
Company confidential
Max. 31
Start
Finish
Input Data
CPICH Scrambling code Ec/Io Measurement position Cell ID, cell position, cell azimuth Cell scrambling code Cell neighbor list
Source
Scanner Scanner Planning Tool Radio Design Planning Tool
Company confidential
Start
Record drive test results with Agilent Scanner in TOP N mode Analysis within Actix Analyzer
Neigbor Wondow
Finish
Time
SC
216 349 83 436
Site
Ewer Park Stud Reading Rubgy FC WOODLEY ATE TILEHURST ATE
Sector
00161126C 00179719B2 00001085A 00001081B
Num Of Samples
97 97 96 90
Company confidential
Numerical Analysis
Start
Run Actix Analyzer Routine
Company confidential
No
Does list include all cells belonging to that site and the first perimeter of cells
Neighbours
Finish
Company confidential
Company confidential
What it shows measured CPICH RSCP level (in dBm) of the strongest serving cell measured CPICH Ec/No level (in dB) of the strongest serving cell
Summary
Any area with RSCP of less than this design threshold is considered a weak coverage area. The RF Planning Guidelines requirement for in-car RSCP for AMR voice is -100 / -99 dBm. (Vendor specific) CPICH Ec/No is the most important measurement in WCDMA for network planning and optimization purposes. It is theoretically equal to RSCP divided by the total wideband received power (RSCP/RSSI). Low Ec/No is the result of interference and poor dominance. Because the measurements were made under unloaded conditions, they represent an ideal, best-case condition. Under fully loaded conditions, observed pilot Ec/No levels would be lower than the unloaded measurements. Any coverage holes, which exist for the unloaded measurements will be worse once the system matures and becomes loaded. Cells that cover a much bigger (or much smaller) area compared to its neighbors. Cells that overshoot beyond the target service area. This plot gives an idea about how well or how bad the coverage of a certain cell is contained. Overshooting cells. Cells with very low RSCP or Ec/No within its target service area. SC clashes identical primary scrambling code being used by cells without sufficient spatial isolation. Strong 4th best server RSCP, say greater than -95 dBm a potential pilot polluter. Number of strong SC > 3. This plot is another way of looking for excessive number of strong pilots. An RSCP threshold of say -95 dBm may be defined. Areas where the UE is transmitting at a higher power due to increased uplink interference, UE issues, hardware issues related to NodeB, TMA, antenna systems. Identify areas where soft handoff / inter frequency or IRAT hand-downs are occurring.
Best server SC
measured CPICH RSCP and Ec/No level (in dBm and dB) for a specified scrambling code measured CPICH RSCP level (in dBm) of the 4th strongest serving cell number of detected pilots above a defined RSCP threshold Measured UE Tx Power level (in dbm)
UE Tx Power
Company confidential
This part of post-analysis deals with the events that were logged through the handsets or UEs. These are: the dropped calls call setup failures. Some of these may be the net effect of problems discussed in the previous section, while some may be due to other reasons. Each event needs to be investigated thoroughly, analyzed and eliminated.
Company confidential
The causes for the failures can be classified into three groups: RF planning-related, network-related tool-related.
Technically, job of the RF engineer may be limited to fixing only those that fall in the first category but the rest still need to be properly documented and referred to the concerned departments for their action.
Company confidential
RF Planning problems
As the name implies, RF planning-related faults have something to do with the radio environment and the related cell / RNC parameters. A myriad of reasons can be the culprit but the usual suspects are:
Inadequate downlink coverage, poor dominance, pilot pollution chances are, many of these faults are directly related to the same problems detected during the Layer 1 analysis Inadequate uplink coverage, uplink interference Uplink power control is responsible for keeping the uplink SIR within target by increasing the UE transmit power when needed. Areas where RSCP is good but UE Tx power is high (e.g. >10 dBm) are signs of radio problems in the uplink path, such as issues with the Node B Rx. The installation of an MHA may be one of the possible solutions to attain good uplink coverage. On the other hand, uplink interference may be the issue. This is easily read and verified from the UE messages. Missing or one-way neighbors a more critical error in WCDMA than in GSM, as they almost inevitably lead to dropped calls as soon as the pilot of the missing neighbor dominates. Rapid field drop e.g. around building corners or natural obstructions. Slow handovers caused by inappropriate handover window parameters or neighbor lists that are too long. Handover speed is directly affected by the handover trigger parameters based on pilot Ec/No for neighbor addition, removal or replacement. Scrambling Code interference Can be resolved by changing the cell primary scrambling code
Company confidential
Network problems
Network-related faults can be due to: Outages and hardware faults in Node B, transmission, RNC or part thereof, including sleeping cells and T1 Issues Core Network faults caused by board / software failures in MSC, SGSN or other core network entities Insufficient hardware caused by lack of hardware resources such as Channel elements or radios Installation faults crossed or wrongly terminated feeders. One of the most difficult to detect are cases where two Rx feeders were accidentally swapped across different sectors. Diversity alarms and TX/RX imbalance scripts should identify these problems. System software bugs RNC trace data can provide more clues for this type of problem Operational intervention e.g. when an active serving cell is manually reset or blocked causing call failures and/or call drops
Alarm logs coming from the NMC will be valuable reference for these networkrelated failures. They must be regularly gathered during the cluster tuning period and should contain the status of each site in the cluster driven, any service affecting alarms or other minor/major/critical alarms and T1 error information
Company confidential
Tool Problems
Sometimes the tools that were used during the measurement or post-processing also contribute to the errors: Software bugs causing either the handset or the data collection program to momentarily hang during the test in some tools, the so-called flat line measurement display is an indication of this phenomenon (eg TEMS) Incorrect tool settings as to how a fault is pegged for instance, some call drops shown by the tool may not be real drops. This can occur both for the drive test collection and post-processing tool Miscellaneous faults like loose connectors, broken rf lines, discharged batteries, etc Incorrect Attenuation settings could be the cause of poor UL / DL coverage and other call related issues
Company confidential
Call Events
Troubleshooting call events takes a more in-depth approach, involving more than just checking the primary measurements. The Layer 3 messages have to be sifted through as well. This is where a thorough understanding of call flow and signaling becomes essential.
To be successful, the Engineer should be well-versed about the UE-UTRAN-CN signaling that takes place in different scenarios. Some important scenarios are listed below. Not all of these events will be encountered during the tests, but it is still worth being able to identify them if they do occur.
AMR MOC RRC and RAB connection establishment SHO soft and softer handovers, measurement reporting events 1a, 1b, 1c, 1d, 1e, 1f, 2d, 2f, 3a
Other issues
During optimization only one parameter per time should be changed. In this way it is possible to see the influence of that particular parameter on the network.
Another important issue is to keep track of all the changes implemented, in order to restore the network to the previous config, if the changes implemented had some undesirable side effects.
Company confidential
Exercise 1
Company confidential
Exercise 2
Display histogram distribution for RSCP Display histogram distribution for Ec/No
Company confidential
Exercise 3
Company confidential
Exercise 4
Company confidential
Exercise 6
Company confidential
Company confidential
Company confidential