RNP Model Tuning Guide
INTERNAL
Product name RNP Product version 1.2
Confidentiality level INTERANL Total 85 pages
RNP Model Tuning Guide (For internal use only)
Prepared by Reviewed by Reviewed by Approved by
Date Date Date Date
Zang Liang
2008-11-17
Huawei Technologies Co., Ltd. All Rights Reserved
2008-12-10
All rights reserved
Page1 , Total85
RNP Model Tuning Guide
INTERNAL
Revision Records Date 2008-1205 2008-1216 2009-3-25
Version Description
Reviewer
1.0
Initial transmittal
Li Peng, Lu Peng, Zang Liang Yangfan
1.1
Correct some errors for easy understanding.
Qin Yan
1.2
Update the operating environment of SPM Tuning to U-Net 2.2.1
Chen Fazhi
2008-12-10
All rights reserved
Author
Page2 , Total85
RNP Model Tuning Guide
INTERNAL
Contents 2.1 About Model Tuning.................................................................................................................................................9 2.2 Flow for Model Tuning...........................................................................................................................................10 3.1 About SPM.............................................................................................................................................................12 3.1.1 Basic Formula...................................................................................................................................................12 3.1.2 Distance and Visibility Between Tx Antenna and Rx Antenna .......................................................................12 3.1.3 Effective Height of Tx Antenna........................................................................................................................13 3.1.4 Effective Height of Rx Antenna.......................................................................................................................14 3.1.5 LOS Amendment for Mountainous Regions....................................................................................................14 3.1.6 Calculating Diffraction Loss.............................................................................................................................14 3.1.7 Clutter loss........................................................................................................................................................14 3.2 Procedure for Tuning SPM.....................................................................................................................................15 3.2.1 Setting Up a Model Tuning Project..................................................................................................................16 3.2.1 Setting Up Propagation Model.........................................................................................................................20 3.2.2 Setting Transmitter...........................................................................................................................................25 3.2.3 Importing and Adjusting Data..........................................................................................................................31 3.2.4 Model Tuning...................................................................................................................................................46 3.2.5 Proposals on SPM Tuning................................................................................................................................51 3.2.6 Analyzing Result and Verifying Model............................................................................................................54 4.1 Configuring Parameters of Volcano Model............................................................................................................59 4.1.1 Configuring Parameters of Volcano Macrocell Model.....................................................................................60 4.1.1 Configure Parameters of Volcano Microcell Model.........................................................................................67 4.1.2 Configuring Parameters of Volcano Minicell Model........................................................................................74 4.2 Tuning Volcano Models..........................................................................................................................................82 4.2.1 Tuning Process..................................................................................................................................................83 4.2.1 Checking and Analyzing Tuning Result...........................................................................................................84
2008-12-10
All rights reserved
Page3 , Total85
INTERNAL
Figures Model Tuning Flow.........................................................................11 Four weighting methods for calculating clutter loss in SPM..............15 Importing data of Clutter class.......................................................17 Importing data of altitude .............................................................18 Imports data of clutter heights.......................................................18 Importing data of vectors (1)..........................................................19 Importing data of vectors (2)..........................................................19 Antenna properties........................................................................20 Importing antenna file....................................................................20 Setting up propagation model........................................................21 Properties of SPM..........................................................................22 Setting SPM parameters.................................................................22 Configuring parameters of Clutter tab for SPM................................24 Configuring parameters of Calibration tab for SPM..........................25 Importing head file........................................................................27 Global parameters of transmitters..................................................27 Configuring transmitter propagation models ..................................28 Setting up new site........................................................................29 Setting up transmitter...................................................................29 Properties of new transmitter.........................................................30 Configuring pilot power..................................................................31 Setting up CW measurements.........................................................32 New CW measurement path............................................................33 Interface after importing data........................................................33
2008-12-10
All rights reserved
Page4 , Total85
INTERNAL
Importing CW measurement...........................................................34 Importing a text file.......................................................................34 General tab displayed after importing CW measurement data..........35 Changing measurement line...........................................................36 CW measurement setup.................................................................36 Setting CW measurement manually.................................................37 Filtering measurement data...........................................................38 Displaying measurement data........................................................39 Properties of the file for tuning coordination system.......................40 Before translating location in map..................................................41 Original coordinates of reference point...........................................42 After translating location in map....................................................43 Coordinates of reference point after translation..............................44 Frequency Difference Setting.........................................................45 Parameters of SPM to be tuned.......................................................47 Setting filtering conditions for SPM tuning......................................48 Tuning SPM (1)..............................................................................49 Tuning SPM (2)..............................................................................49 Automatic Calibration....................................................................51 Select parameters for calibration....................................................51 Statistics report after tuning..........................................................55 Measurement parameters of tuned model.......................................56 Comparison curve..........................................................................57 Properties window of measurement data of tuned model.................57 Error distribution...........................................................................58 Volcano models displayed in U-Net..................................................60 Parameters in General Tab for Volcano Macrocell model...................61 Parameters in Map data tab for Volcano Macrocell model.................62 Parameters in Clutters Tab for Volcano Macrocell model...................63 Parameters in vectors tab for Volcano Macrocell model....................64
2008-12-10
All rights reserved
Page5 , Total85
INTERNAL
Parameters in Parameter tab for Volcano Macrocell model...............65 Parameters in Tuning tab for Volcano Macrocell model.....................66 Parameters in General Tab for Volcano Microcell model....................67 Parameters in Map data tab for Volcano Microcell model..................68 Parameters in Clutters Tab for Volcano Microcell model....................69 Parameters in vectors tab for Volcano Microcell model.....................70 Correct configuration of vectors of building type.............................71 Menu file for vector map................................................................72 Parameters in Ray Tracing tab for Volcano Microcell model...............72 Parameters in Parameter tab for Volcano Macrocell model...............73 Parameters in General Tab for Volcano Miniocell model....................75 Parameters in Map data tab for Volcano Microcell model..................76 Parameters in Clutters Tab for Volcano Minicell model......................77 Parameters in vectors tab for Volcano Minicell model.......................78 Correct configuration of vectors of building type.............................79 Menu file for vector map................................................................80 Parameters in Ray Tracing tab for Volcano Minicell model.................80 Parameters in Parameter tab for Volcano Minicell model..................81 Volcano tuning dialog box..............................................................83 Selecting automatic tuning mode for Volcano Microcell model..........84 Tuning report.................................................................................84 Calibration result...........................................................................85
2008-12-10
All rights reserved
Page6 , Total85
INTERNAL
2008-12-10
All rights reserved
Page7 , Total85
INTERNAL
Tables Frequency differences for GSM900 in different propagation environments................................................................................45 Frequency differences for DCS1800 in different propagation environments................................................................................45 Default values of SPM coefficients..................................................46 Value range of K parameters..........................................................50 Typical values of clutter losses.......................................................52
RNP Model Tuning Guide Key words Model Tuning, SPM, Volcano Model
Abstract This guide introduces flow and principle for model tuning. The operation procedures for SPM and Volcano model tuning are described based on U-NET tool. There are some rules for guaranteeing correct tuning results. The U-NET tool software version is U-NET 2.2.1(Build 2613) in this book.
Acronyms and abbreviations: Acronyms and abbreviations
Full Spelling
CW SPM LOS/NLOS
Continuous Wave standard propagation model Line of sight/No Line of sight
2008-12-10
All rights reserved
Page8 , Total85
INTERNAL
1
Overview
The propagation model lays a foundation for the cell planning of a mobile communication network. The accuracy of the propagation model determines whether the cell planning is reasonable and whether operators can meet the requirements of users through cost-effective and rational investments. Therefore, it is necessary to tune the propagation model in order to obtain the wireless propagation model that complies with the actual environment of the specific area, improve the accuracy of the coverage prediction and lay a solid foundation for network planning. This document introduces rationales for propagation model tuning, methods and principles for propagation model tuning based on the U-Net tool. The methods and principles include method and flow for processing the data obtained in the CW test and tuning the propagation model by using the data in the U-Net tool and some rules for guaranteeing correct tuning results. This guide includes main chapters as below. Chapter 2: Introduce the flow and basic knowledge for model tuning. Chapter 3: Describe the process for SPM model tuning in a detail. Chapter 4: Introduce how to tune Volcano model.
2
Flow for Model Tuning
2.1 About Model Tuning There are two propagation model research methods: 2008-12-10
All rights reserved
Page9 , Total85
INTERNAL −
Theoretical analysis method based on radio propagation
Actual measurement and statistics method based on large amount of test data and empirical formulae −
The radial tracking model integrated into the planning software, which can be put into commercial use, such as Volcano model, WaveSight model and WinProp model, and they are the representations of propagation model research through the theoretical analysis method, but this type of model requires a high precision (at least 5 m precision), including 3D digital map of the building information. Accuracy of model prediction is closely related to the precision and accuracy of the digital map. Those factors that affect the propagation of radio signals such as moving vehicles cannot be considered in the current theoretical analysis method, but the general theoretical analysis method requires some approximation and simplification of the propagation environment, thus causing certain errors. At present, the propagation model based on the theoretical analysis method has not been applied in large scale. In the actual measurement and statistics method for the propagation model, the most famous statistics model is Okumura model. This model is a propagation model represented by curves and was built by Okumura based on a large amount of test data collected in Japan. On the basis of the Okumura model, the regression method is used to fit out resolution empirical formulae to facilitate computation. These empirical formulae include the Hata model formula, which is applicable to the GSM900 macro cell, and the COST 231-Hata formula, which is applicable to the GSM1800 macro cell. These empirical formulae also include the COST 231 Walfish-Ikegami model formula, which is applicable to the microcell, and the Keenan-Motley formula, which is used in the indoor propagation environment. These formulae are complex and errors may occur as against the actual environment.
2.2 Flow for Model Tuning During the actual field strength predication, the tuned Hata model is used as the prototype, and the planning software is used to import the data that is obtained from the radio propagation features test on the local radio environment. Tune the coefficients in the model formula and then the propagation model for actual predication is obtained. 2.2 shows the flow for model tuning.
2008-12-10
All rights reserved
Page10 , Total85
INTERNAL
Figure 1.1 Model Tuning Flow
The procedure of “Tune model” in 2.2 is as below. 1. Select a model and set parameter values. The values can be default values on this frequency or that of tuned parameters similar to the terrain in other places. 2. Perform radio propagation predication by using the selected model, and compare the predicated value with the drive test data to obtain a difference. 3. Change the model parameters based on the obtained difference. 4. Perform constant iteration and processing till the mean square error and the standard deviation between the predicated value and the drive test data are minimized. Then the parameter values of the model obtained after tuning are what we need.
2008-12-10
All rights reserved
Page11 , Total85
INTERNAL
3
SPM Tuning
3.1 About SPM SPM is based on the formula of Cost231-Hata model. Compared with Cost231-Hata, SPM has the following new features:
The factors are variable.
The diffraction on clutter is added.
SPM supports using different constant K1 and distance coefficient K2 for LOS/NLOS and near/far region.
Due to the previous new features, SPM is more flexible and applies to more scenarios. You can tune SPM according to the data of CW measurement, namely, the adjustment of parameters.
3.1.1 Basic Formula SPM is based on the following formula:
Lmod el = K 1 + K 2 log ( d ) + K 3 log ( HTxeff ) + K 4 × Diffractio n loss + K 5 log ( d ) × log ( HTxeff ) + K 6 ( H Rxeff ) + K clutter f ( c Wherein,
K1: a constant (dB), related to frequency.
K2: the multiplier (distance factor). It shows how the field strength changes as the distance changes.
d: the horizontal distance (m) between the Tx antenna and Rx antenna.
K3: the multiplier of log(HTxeff). It represents the variation of field strength as the height of Tx antenna changes.
HTxeff: effective height of Tx antenna (m)
K4: multiplier of diffraction loss. It represents the strength of diffraction.
Diffraction loss: diffraction loss due to obstacles (dB).
K5: multiplier of log(HTxeff)log(d).
K6: multiplier of H Rxeff . It represents the variation of field strength as the height of Rx antenna changes.
H Rxeff
: Effective height of Rx antenna (m)
Kclutter: multiplier of f(clutter). It is the weighting factor of clutter loss.
f(clutter): weighted-average loss due to clutter.
3.1.2 Distance and Visibility Between Tx Antenna and Rx Antenna In each calculation, SPM uses the following aspects. 2008-12-10
All rights reserved
Page12 , Total85
INTERNAL
2
Distance Between Tx Antenna and Rx Antenna If the distance between Tx antenna and Rx antenna is shorter than the maximum distance defined by the operator, the Rx antenna is considered as near the Tx antenna. SPM will use the parameters marked with Near transmitter for calculation. If the distance between Tx antenna and Rx antenna is longer than the maximum distance defined by the operator, the Rx antenna is considered as far from the Tx antenna. SPM will use the parameters marked with Far from transmitter for calculation.
3
Visibility According to terrain and clutter high, SPM judges whether the receiver is in the light of sight (LOS) range. If you do not use the clutter height layer, SPM calculates LOS with the terrain height map only. If you use the clutter height layer, SPM calculates LOS with the terrain and clutter height maps. If the receiver is in the sight of LOS, SPM use (K1,K2)LOS; otherwise, SPM uses (K1,K2)NLOS.
3.1.3 Effective Height of Tx Antenna There are six methods to calculate effective height of Tx antenna H txeff
as below:
Height above ground The height above ground is the height of Tx antenna above ground.
Height above average profile Determining the height of Tx antenna depends on average ground height, which is calculated on the lateral section where the transmitter and receiver are.
Slope at receiver between 0 and distance min Calculate the height of Tx antenna with the slope of the ground where the receiver is.
Spot Ht
Abs Spot Ht
Enhanced slope at receiver
U-Net supports a new method to calculate effective height of Tx antenna, called "Enhanced slope at receiver". The methods of "1-Height above average profile" and "0-Height above ground" apply to plain region while other methods are for mountainous regions. This does not mean that the methods for mountainous regions do not apply to plain regions. The best method is to adjust these parameters and to produce a most suitable tuning result.
2008-12-10
All rights reserved
Page13 , Total85
INTERNAL
3.1.4 Effective Height of Rx Antenna H Rxeff = ( H Rx + H 0Rx ) − H 0Tx Wherein,
H Rx : the receiver antenna height above ground (m).
H 0 Rx : the ground height (ground elevation) above sea level at the receiver (m).
H 0Tx : the ground height (ground elevation) above sea level at the transmitter (m).
The calculation of effective heights of antennas H Rxeff height data is not realized, the calculation will fail.
and H Txeff
is based on the DTM lateral section. If importing
3.1.5 LOS Amendment for Mountainous Regions An optional amendment condition is that SPM can amend path loss of mountainous regions on the condition that the transmitter and receiver are LOS.
3.1.6 Calculating Diffraction Loss U-Net calculates diffraction loss on the lateral section of transmitter and receiver with the following four methods:
Deygout
Epstein-Peterson
Deygout with correction
Millington
For the urban areas or the urban areas with rural areas, the operator can use Deygout and EpsteinPeterson (these two methods also apply to mountainous regions).
3.1.7 Clutter loss U-Net calculates the maximum distance f(clutter) from the receiver as below: n
f ( clutter ) = ∑ Li w i i =1
Wherein,
L: clutter loss defined by the operator in the Clutter tab
w: the weighting factor for applying weighting function
n: number of spots to be considered in the lateral section. These spots are distributed according to the accuracy of lateral section.
There are four weighting functions as below:
2008-12-10
Uniform weighting function: w i =
1 n
All rights reserved
Page14 , Total85
INTERNAL
Triangular weighting function:
wi =
di n
∑d j =1
j
d i = D − d i' . The d'i is the distance between the receiver and the ith spot. The D is the maximum defined distance.
d log i + 1 D Logarithmic weighting function: w i = n dj log + 1 ∑ j =1 D
Exponential weighting function
wi =
e n
di D
∑e
−1 dj D
−1
j =1
Figure 1.1 Four weighting methods for calculating clutter loss in SPM w i=f(di) Uniform w eighting function Triangular w eighting function Loragithmic w eighting function Exponential w eighting function
wi
di
3.2 Procedure for Tuning SPM To make the model more applicable for a region, you can tune the model with the data of radio propagation feature test (usually CW measurement), namely, adjusting the parameters of model.
2008-12-10
All rights reserved
Page15 , Total85
INTERNAL
CW data measurement and filtering processing are introduced in the guide Radio Propagation Feature Test. Here shows the process of data input. Before importing head files, proceed as below: Step 2 Set up a new project. Step 3 Import digital maps. Step 4 Set the coordination system. Step 5 Import antenna pattern The contents from step 1 to step 4 are simply described here and can be known detailedly in the UNet Operation Manuel. Step 6 Set up the model to be tuned. ----End After importing head files, compare the related setting values with contents in head files.
3.2.1 Setting Up a Model Tuning Project 2. Creating a New Project In U-Net, click the New shortcut button or select File > New. In the pop-up Project type window, there are several types. They are CDMA2000 1XRTT 1XEV-DO, GSM GPRS EGPRS, IS-95CDMA one, UMTS HSDPA,TD-SCDMA, Microwave Radio Links, WIMAX 802.16d, WIMAX 802.16e.
3. Setting Coordination System In U-Net, select Tools > Options. The Projection is the main coordination system while the Display is the side coordination system. There is a button on the right of Projection window. For digital map of China, select the WGS84 UTM Zones coordination system, and sometimes Asia-Pacific coordination system. This depends on the coordination system on which the digital map is based. The longitude of Zhangzhou City, Fujian Province, China is 114°–120° east,
If you select WGS84 UTM Zones coordination system, so you shall select WGS84/UTM zones 50N.
If you select Asia-Pacific coordination system, so you shall select Beijing 1954/GaussKruger 20N.
The previous two coordination systems apply to the digital maps for Zhangzhou. The side coordination system depends on the main coordination system, namely,
Main coordination system: WGS84 / UTM zone 50N Side coordination system: WGS84
Main coordination system: Beijing 1954/Gauss-Kruger 20N Side coordination system: Beijing 1954 or WGS84
Identify the coordination system of the current map as below: In the height directory of digital map, there is a file named projection. There are three lines. The first line is the ellipsoid for projection (such as WGS84 and Russian Krassowsky). The section line is the projection mode. The third line is the central meridian and coordination system shift. Find the coordination system corresponding to the projection in U-Net. Pay attention to the difference of coordination system of The Northern Hemisphere and The Southern Hemisphere. The coordination 2008-12-10
All rights reserved
Page16 , Total85
INTERNAL
system for The Northern Hemisphere is usually marked with the letter N.
4. Importing Digital Map U-Net can use the files of various formats without conversion. It supports the following formats: Raster data: DEM, terrain distribution data, traffic data, scanned maps. The formats of scanned maps include BIL, TIF, BMP, MSI Planet, and original binary files. −
Vector data: MSI Planet, DXF, MIP-Mapinfo, and Arcview Shapefile.
−
The following paragraphs describe how to import digital maps that are commonly used by Huawei. For importing other types, see U-Net Usage Guide.
Clutter Map
Read the index file according to the saving path (usually clutter or DLU) of clutter map files. Select Clutter class as shown in 3.2.1, and then click OK. Figure 1.1 Importing data of Clutter class
Height Map
Read the index file in the saving path (usually height or DTM) where the height map file is saved. As shown in 3.2.1, in Data type box, select Altitude, and then click OK.
2008-12-10
All rights reserved
Page17 , Total85
INTERNAL
Figure 1.2 Importing data of altitude
Clutter Height Map
Read the index file in the saving path (usually Building or DHM) where the height map file is saved. As shown in 3.2.1, in Data type box, select Clutter Heights, and then click OK. Figure 1.3 Imports data of clutter heights
Vector Map
Read the index file in the saving path (usually Vector) where the vector map file is saved. As shown in 3.2.1, in Data type box, select Vectors, and then click OK.
2008-12-10
All rights reserved
Page18 , Total85
INTERNAL
Figure 1.4 Importing data of vectors (1)
Figure 1.5 Importing data of vectors (2)
By select the Embed in document right under the Geo drop-down list, you can choose to display one or four types of maps. For the operations like modifying the properties of map, see the corresponding manual. Not all maps include the four data types. Import the corresponding maps according to the map conditions and project's requests in actual operations.
5. Importing Antenna Information In the Explorer pane, click Data tab. Select Antennas > New. In the popup Antennas new elements properties window, click General, and input the parameters like antenna type, vendor, and gain. In the Horizontal pattern and Vertical pattern, import the fading table for the antenna. Namely, copy all the data in the Excel and paste it to the corresponding table. 2008-12-10
All rights reserved
Page19 , Total85
INTERNAL
Figure 1.1 Antenna properties
For the properties of antenna, see the U-Net User Manual. If the current antenna is present in the original antenna library, you do not need to re-import and you can use it directly. Import the corresponding antenna file directly if there is. Select File > Import. In the File dialog box, change the file type to "Planet Database". In the Planet data to be imported dialog box, import the antenna index file (named index) in the Antenna box. Click OK, and then click OK. You can see the imported antenna file in the Explorer window of U-Net. Figure 1.2 Importing antenna file
3.2.1 Setting Up Propagation Model In the Explorer pane, click Modules tab. Right-click Standard propagation model, and select Duplicate to create a new model named copy of standard propagation model. You can define as you wanted based on this template. You need to set up models as many as the models to be tuned in 2008-12-10
All rights reserved
Page20 , Total85
INTERNAL
an area. Figure 1.3 Setting up propagation model
1. Configuring Parameters in the General Tab In the Explorer pane, click Modules tab. Double click the new Copy of Standard propagation model. In the popup dialog box, click General tab, fill the name for the model to be tuned, such as shanghaiCWtest, as shown in 3.2.1.
2008-12-10
All rights reserved
Page21 , Total85
INTERNAL
Figure 1.4 Properties of SPM
6. Configuring Parameters in the Parameters Tab In the Copy of Standard propagation model window, as shown in 3.2.1, there are default values for each parameter. Figure 1.1 Setting SPM parameters
2008-12-10
All rights reserved
Page22 , Total85
INTERNAL
Before tuning propagation models, you shall configure the parameters in the Parameters tab, detailed as below:
Near Transmitter\Max. Distance It is 0 by default no matter it is near or far.
Near Transmitter & Far from Transmitter K1 is 17.4 and K2 is 44.9 by default. The initial parameters for LOS and NLOS are the same.
Effective Antenna Height Method: By default, if the overall terrain of target area is flat without great undulation, you are recommended to select 1-height above average profile. If the overall terrain of target area is with great undulation (with a fall of 50 m or above), you are recommended to select 5-enchanced slope at receiver. The Distance min (m) and Distance max (m) do not serve in calculation in moduel tuning, so use the default values. Use the default value of K3.
Diffraction Method: select 1-Deygout by default. K4: If there is not height information about clutter in the map and there is no great undulation in the area, you are not recommended to adjust K4 and you can configure K4 to 0; otherwise, configure it to 1.
Other parameters K5: use the default value. K6: use the default value. Kclutter: you can configure Kclutter to 0 in tuning; namely, you do not count clutter loss. The CW test usually proceeds in outdoor open land, so there are inadequate spots. As a result, the clutter loss is not adjusted according to recommendation. The default value of clutter loss serves in simulation forecast, so the default Kclutter is 1. Other parameters /hilly terrain correction: configure it to 1-yes only when the total terrain is with great undulation (with a fall over 50 m); otherwise, configure it to 0-no. Profile: use the default value. Grid calculation: use the default value.
2008-12-10
All rights reserved
Page23 , Total85
INTERNAL
7. Configuring the Parameters of Clutter Tab Figure 1.1 Configuring parameters of Clutter tab for SPM
Parameters are configured as below:
Height\User clutter heights: it refers to whether to consider the clutter height in calculating diffraction loss when the imported digital map contains clutter height data.
If you use 2D digital map (20-meter solution or lower) without clutter height layer, you can configure Use clutter heights to 0-No. −
If you use 3D digital map (5-meter solution or higher) with clutter height layer and you want to calculate diffraction loss with clutter height, you can configure Use clutter heights to 1-yes. −
If you use 3D digital map (5-meter solution or higher) with clutter height layer and you do not want to calculate diffraction loss with clutter height, you can configure Use clutter heights to 0-No. −
As previously described, if you configure Use clutter heights to 1, namely, you want to calculate the diffraction with clutter height, you need configure K4 to 1, the default value.
Max.distance: Configure to 0 by default.
Losses per clutter class: it refers to clutter losses. The CW test spots are all in open
2008-12-10
All rights reserved
Page24 , Total85
INTERNAL
land, so you must configure all the clutter losses to 0 in model tuning. Do not adjust clutter loss in model tuning. However, you need configure clutter loss to the recommended value in simulation forecast.
Other parameters: use the default values.
8. Configuring the Parameters of Calibration Lable 3.2.1 shows the Calibration tab. Figure 1.1 Configuring parameters of Calibration tab for SPM
When the test data is already imported, the test paths will be displayed in the CW measurement path(s) to be used box. LOS and NLOS: If you do not want to distinguish LOS and NLOS, you need select two buttons. If you want to tune LOS or NLOS parameters respectively, you can select them respectively for tuning.
3.2.2 Setting Transmitter Configure the site and transmitter as usual. Set up a transmitter for CW test. Pay attention to the following aspects:
Longitude and latitude of site
Transmit power at feeder port
Losses on feeders and connectors
Antenna type
You can set the transmitter manually or by importing head file. 2008-12-10
All rights reserved
Page25 , Total85
INTERNAL
9. Importing Head File Creating a head file The suffix of head file is hd. A head file corresponds to a test site, so you shall make several head files for several sites. Every two items are separated by a blank. The following describe an example. test.dat DATE test 741790 32 54 0 0 0 23.04499751 113.7509966 Survey 0 GPS 0 0 0 0 hard Wherein,
test.dat: the name of the file that saves test data for the site. If there are multiple test files for a site, it is better to combine them into one. Now the data in the file can not be automatically imported. Therefore you can delete it and re-import the data files when test.
DATE: the test data. You can configure it to DATE.
Test: the name of test site. You can change it accordingly. It is displayed after successful import.
741790: the type of antenna used by the site. The antenna type is available before import in U-Net. If it is unavailable, the U-Net automatically sets up the antenna type with the same name. the pattern and gain are by default, so you need modify them accordingly. It is recommended that you set up the corresponding antenna type before importing the head file.
32: the effective height of antenna on the site.
54: the transmit power of feeder port. When you configure it in U-Net, you can obtain pilot power of NodeB by deducting antenna gain from EIRP. It is recommended as below: Transmit power of feeder port = output power of transmitter - total loss on the Tx feeder - loss on feeder connectors + antenna gain + Rx antenna gain - total Rx loss. You shall guarantee that the length of Tx feeder, the gain of Rx antenna, and Rx loss are configured to 0. They are 0 by default.
0 0 0: the three 0's are the azimuth, down tilt, and squire size respectively. They are 0 for omnidirectional antenna.
23.04499751: the northern latitude degree of test site. Set it accordingly.
113.7509966: the eastern longitude degree of test site. Set it accordingly. Note that the latitude degree is before the longitude degree. If it is southern latitude or western longitude, put a minus symbol before the value.
Survey 0 GPS 0 0 0 0 hard: the last eight parameters of head file. They can be fixed. The 1, 2, 3, 4, 6, 10, and 11 items are usually fixed. The method for obtaining the template for head file the same as that for manual, and you can also make it.
When the head file is imported, the inside antenna type is not set up yet. The gain of the antenna of the type automatically set up by U-Net is 0. As a result, the pilot power is the same as the transmit power of feeder port, because the antenna gain is not automatically deducted. After the antenna pattern and gain are changed, U-Net judges that the transmit power of feeder port is higher than actual power, so error occurs. Therefore you need set up the antenna type before importing head file, and then check whether the pilot power equals to the transmit power of feeder model minus antenna gain.
Importing a head file 2008-12-10
All rights reserved
Page26 , Total85
INTERNAL
Step 1 In U-Net, click the Explorer pane, click the Data label, and right-click CW Measurement (or select File > Import). In the drop-down list, select Import. Step 2 A window is displayed, as shown in 3.2.2. Select the target head file and select its file type as *.hd, and then open the file. Figure 1.1 Importing head file
Step 3 In U-Net, click the Explorer pane, click the Data label, and right-click transmitter. Select properties in the window, and the Transmitters properties window is displayed. Click the Global parameter tab, as shown in 3.2.2. Configure frequency and keep the first carrier unchanged as shown in 3.2.2. Figure 1.1 Global parameters of transmitters
Step 4 Click the Propagation tab, as shown in 3.2.2.
2008-12-10
All rights reserved
Page27 , Total85
INTERNAL
Figure 1.1 Configuring transmitter propagation models
Select the set model in the Propagation model drop-down list respectively. When there are multiple sectors, you can set the model for each sector. You can configure the radius and resolution respectively. The radius is the maximum distance from the test data to the site in model tuning. Resolution is the accuracy of map. You do not need configure other parameters.
10. Manual Setup Step 1 Set up a site first. Right-click Sites in the navigation tree, and select New. A properties dialog box is displayed, as shown in 3.2.2. Input the site name, longitude, and latitude. Since only the model tuning is necessary here, you do not have to configure the parameters in the Equipment tab.
2008-12-10
All rights reserved
Page28 , Total85
INTERNAL
Figure 1.1 Setting up new site
Step 2 Set a Transmitter. In the similar way, you can configure parameters in the General tab on the Transmitters new element properties. Input the name and site. The Dx and Dy are usually configured to 0 m, because the location for recording data is the location of antenna. Figure 1.1 Setting up transmitter
Step 3 In the Transmitter tab, input the losses of Tx feeder, antenna model, height, down tilts, as shown in 3.2.2. Click OK.
2008-12-10
All rights reserved
Page29 , Total85
INTERNAL
Figure 1.1 Properties of new transmitter
Step 4 In the navigation tree, right-click new Transmitter, and select Properties dialog box. Select the Cell tab, and configure pilot power.
2008-12-10
All rights reserved
Page30 , Total85
INTERNAL
Figure 1.1 Configuring pilot power
Only the pilot power is used in model tuning, so other configurations have no impact on the tuning result. The pilot power minus loss of Tx feeder is the transmit power of feeder port. The transmit power of feeder port plus antenna gain is EIRP. You can configure the previous loss of Tx feeder to 0, and the pilot power shall deduct loss of Tx feeder. −
Some tests provide EIRP (such as there is only one head file *.hd), so you can configure the loss of Tx feeder to 0 and configure the pilot power to EIRP minus antenna gain so that the transmit power of feeder power is correct. −
Other configurations, such as frequency and propagation mode, are the same as importing files. You can refer to the previous section.
3.2.3 Importing and Adjusting Data 11. Organizing Data of CW Measurement U-Net needs longitude, latitude, and field strength of signal. U-Net supports the data of multiple formats. The data of a site is usually combined into a text file, and then import the file or paste directly. The following sections describe the process.
12. Importing DT Data There are two methods to import DT data.
1）Paste Step 1 In the Explorer pane of U-Net, select the Data tab, right-click CW measurements, and select New in the menu.
2008-12-10
All rights reserved
Page31 , Total85
INTERNAL
Figure 1.1 Setting up CW measurements
Step 2 Configuring the parameters shown in 3.2.3 and then click OK. −
Input the name of test file in the Name text box. You can name the test file
accordingly. In the Transmitter area, input the name of site in the Name drop-down list and the frequency in the Frequency spin box. −
In the Receiver area, the Height is the height of test antenna. The Rx antenna is usually mounted on the roof of vehicle, so the height depends on the height of vehicle. The default height is 1.5 m. The gain and loss are usually 0. The value is already considered in the transmit power of feeder port (note: the antenna gain and feeder loss of DTI is 4 dB, so they can counteract each other). −
−
In the Coordinates area, select the correct coordinates, otherwise serious error may
happen. In the Measurements area, the unit is dBm by default. The X, Y, and M columns are longitude, latitude, and measured level respectively. After you copy the DT data in an Excel table, you can import the data into U-Net database by clicking the Paste button −
2008-12-10
All rights reserved
Page32 , Total85
INTERNAL
Figure 1.1 New CW measurement path
Figure 1.2 Interface after importing data
2）Importing DT File Step 1 In the Explorer pane of U-Net, select the Data tab, right-click CW measurements, and select Import in the menu. 2008-12-10
All rights reserved
Page33 , Total85
INTERNAL
Figure 1.1 Importing CW measurement
Step 2 In the pop-up box, select the DT file. 3.2.3 shows an example of importing a text file. Figure 1.1 Importing a text file
Step 3 In the General tab, define the parameters in the same way as the previous method.
2008-12-10
All rights reserved
Page34 , Total85
INTERNAL
Figure 1.1 General tab displayed after importing CW measurement data
Step 4 Configuring parameters in the Setup tab as shown in 3.2.3. Configuration drop-down list involves the combination mode, sorting mode, and filtering standards. After you set the combination, sorting, and filtering standards, you can save the current data structure to Configuration1. No matter how the data structure changes, you can restore the data structure by selecting Configuration1. You can save several Configurations. In actual operations, you can skip this step. −
In the File area, as shown in 3.2.3, the 1st measurement line indicates the line to start with. In 3.2.3, the line starts at the second one, so you input 2 in the 1st measurement line text box. −
2008-12-10
All rights reserved
Page35 , Total85
INTERNAL
Figure 1.1 Changing measurement line
Step 5 After you click the Setup button shown in 3.2.3, a CW measurement setup dialog box is displayed, as shown in 3.2.3. For the * drop down list, select X (File) and Y (File) respectively, and then click OK. Figure 1.1 CW measurement setup
2008-12-10
All rights reserved
Page36 , Total85
INTERNAL
This correspondence may also be achieved by clicking on the first line of the tale header shown in. 3.2.3. Usually you only need to match the longitude, latitude and measurement Figure 1.2 Setting CW measurement manually
Step 6 Click the Import button to import data shown in 3.2.3.
13. Configuring Properties of Data You can check the properties of data by opening the properties dialog box after select the data set to import. You can set the filtering conditions in the Parameters properties tab. The filtering conditions include the following aspects:
Distance range: The minimum distance is usually 100 m to 200 m. The principia for configuring the maximum distance is about as large as twice of forecasted cell radius.
Range of field strength: The filtering conditions for signal strength depend on testers. The range is usually –120 dBm to –40 dBm when test with E7476. The range is usually – 110 dBm to – 40 dBm when test with DTI.
Azimuth range: Engineers use omnidirectional antennas in CW measurement, so you do not need filter by antenna azimuth.
Clutter: Filter the water and buildings which the actual test cannot cover. The test route is in the outdoor land, and it may cover some spots that are defined by GPS, so you shall filter these spots.
After you set filtering conditions, select the Delete points outside the filter check box in 3.2.3, and click Apply to delete unnecessary points.
2008-12-10
All rights reserved
Page37 , Total85
INTERNAL
Figure 1.1 Filtering measurement data
When the measurement data is imported, you can change the mode and color to display the data in the properties window. The detailed method is neglected herein. The data is displayed, as shown in 3.2.3.
2008-12-10
All rights reserved
Page38 , Total85
INTERNAL
Figure 1.2 Displaying measurement data
14. Tuning Coordination System When the measurement route is different from the actual road, you must adjust the coordinates of map with the following two methods: −
Adjust the zero point of map
−
Adjust the coordinates of measurement data (recommended)
1）Adjusting the Zero Point of Map
4
Measure the horizontal and vertical deviation values with U-Net.
2008-12-10
All rights reserved
Page39 , Total85
INTERNAL
5
Import files such as domo.b or demo.b in the GEO/clutter classes and Digital Terrain Model.
Right-click domo.b, and select Properties. A domo.b properties window is displayed, as shown in 3.2.3. Figure 1.1 Properties of the file for tuning coordination system
Step 2 According to the deviation direction, deduct the deviations from or add the deviation to the displayed data of X axis and Y axis, and the two deviations shall be the same. You may have to adjust the data for multiple times until the actual route match the test data. You cannot adjust the vector file with this method, so you need re-set up coordination system or adjust the coordinates in the map file, and then re-import the data. Therefore the process is complicated. If you use SPM, the deviation of vector lay has no impact on the accuracy of tuning, so you can neglect it. ----End
2）Adjusting the Coordinates of Measurement Data The coordinate values of CW measurement files are longitude and latitude, so you shall obtain the deviation of longitude and latitude as below: Step 1 Import CW measurement data and a raster map, such as clutter class layer.
2008-12-10
All rights reserved
Page40 , Total85
INTERNAL
Figure 1.1 Before translating location in map
6
Select a reference point, such as the corner of a house. Click the point and record its coordinates
of longitude and latitude (Ax, Ay).
2008-12-10
All rights reserved
Page41 , Total85
INTERNAL
Figure 1.1 Original coordinates of reference point
7
Import the raster map with geocoding and minimize the deviation of CW measurement data and
actual route.
2008-12-10
All rights reserved
Page42 , Total85
INTERNAL
Figure 1.1 After translating location in map
8
Record the coordinates of longitude and latitude (Bx, By) of reference point after translation;
calculate the deviation of them (dx, dy) = (Ax-Bx, Ay-By).
2008-12-10
All rights reserved
Page43 , Total85
INTERNAL
Figure 1.1 Coordinates of reference point after translation
9 10
Copy the CW measurement data to an Excel table. Add the deviation (dx, dy) to all the
coordinates in Excel, and then re-save it to a text file.
Delete the map and CW data, re-import them.
----End The second method is complicated, but it guarantees that the data matches the route. Therefore, it is preferential. This guide describes Volcano model in the following part, so the vector layer is necessary and is imported in the properties dialog box, so the first method does not work and the second one works.
1. Frequency Difference Setting When the 2G test data is used to tune a 3G propagation model, it is necessary to set different path loss differences because there are different path losses from the transmitter to a same drive test point. You can complete the setting on the Frequency and Model tab page in the Prediction Based on 2G DT Data dialog box of the U-Net, as shown in 3.2.3. 2008-12-10
All rights reserved
Page44 , Total85
INTERNAL
Figure 1.1 Frequency Difference Setting
The U-Net provides three default values in editing areas at three frequency bands. Because path losses vary in different propagation environments, 3.2.3 and 3.2.3 provide some frequency differences of GSM900 and DCS1800 in different propagation environments. You can adjust these differences according to your own requirements. Table 1.1 Frequency differences for GSM900 in different propagation environments Propagation Environment
Frequency Difference
Large cities
14.38 dB
Medium cities
11.35 dB
Suburbs
9.02 dB
Quasi-flat rural areas (with certain clutters)
7.34 dB
Flat rural areas
7.34 dB
Table 1.2 Frequency differences for DCS1800 in different propagation environments Propagation Environment
Frequency Difference
Urban areas
1.549 dB
2008-12-10
All rights reserved
Page45 , Total85
INTERNAL
Propagation Environment
Frequency Difference
Suburbs
1.21 dB
Quasi-flat countryside (with certain clutters)
0.95 dB
Flat rural areas
0.95 dB
As shown in 3.2.3, if you select the Auto Setting DT Based Model check box, the U-Net sets the propagation model of the transmitters for the new CW data to DT Based Model after the data is imported; if you do not select the Auto Setting DT Based Model check box, the U-Net sets the propagation model of the transmitters to the default propagation model. DT Based Model is the name of the newly created propagation model.
3.2.4 Model Tuning 2. Manual Model Tuning Step 1 Initial Values of SPM Tuning Tune SPM with the initial values firstly. The default scenario for SPM in U-Net is urban area. The default height of Rx antenna is 1.5 m. when the actual conditions are different, change the initial values accordingly. SPM origins from HATA model, so you can obtain the equivalent coefficients of SPM according to Cost231-Hata model. The equipment coefficients also serve as the default values and initial values for tuning SPM. Table 1.1 Default values of SPM coefficients SPM coefficien ts
Frequency 450 MHz
900 MHz
935 MHz
1805 MHz
2110 MHz
K1
4.3
12.1
12.6
22.0
24.3
K2
44.9
K3
5.83
K4
0.5
K5
–6.55
K6
0
In the downtown areas of large and medium cities, increase K1 by 3 dB for 1805 MHz and 2110 MHz networks. Adjust the K1 accordingly for suburban and rural areas.
Step 2 Check Initial Parameters In the Modules tab, select the configured model to be tuned. Double click it or select its properties by right-clicking it, and an urban properties window is displayed, as shown in 3.2.4.
2008-12-10
All rights reserved
Page46 , Total85
INTERNAL
Figure 1.1 Parameters of SPM to be tuned
According to the method to configuration parameters in 3.2.1, check whether the parameters in each tab are suitable. If they are suitable, click OK. Right-click the model, and select duplicate. Copy a configured model and tune SPM based on the copied model. Step 3 Select Data and Set Filtering Conditions Right click the copied model and select calibration. A Calibration window is displayed, as shown in 3.2.4. Select the data set to be tuned, then select the Assisted Calibration.
2008-12-10
All rights reserved
Page47 , Total85
INTERNAL
Figure 1.1 Setting filtering conditions for SPM tuning
Note: Before doing calibration, you need to set the filtering conditions including distance, field, and LOS/NLOS. Wherein, the distance and field are set in the measurement data properties dialog box in 3.2.3. If you want to tune K1 and K2 respectively according to LOS and NLOS, you can tune model by selecting LOS and NLOS respectively; otherwise, select them simultaneously. For whether it is necessary to tune LOS and NLOS parameters respectively, see 3.2.5LOS/NLOS. Step 4 Tune Model Click the right Calibration button in 3.2.4, and a Calibration window is displayed, as shown in 3.2.4. Select the variable and tune it by clicking Identify button. Once you select a variable, you can see the correlation with the selected variable on the right.
2008-12-10
All rights reserved
Page48 , Total85
INTERNAL
Figure 1.1 Tuning SPM (1)
Click the Identify button so that U-Net tunes the multiplier of the selected variable. You can know the amendment of the variable by checking correction. The relation between variables and K parameters can be known from the basic formula in 3.1.1. The current value of K parameters equals to the initial value plus amendment, as shown in 3.2.4. Figure 1.2 Tuning SPM (2)
The most influential variable for tuning is log(D). You shall tune the multiplier K2 preferentially, and the K1 will be automatically tuned.
2008-12-10
All rights reserved
Page49 , Total85
INTERNAL
After tuning a parameter, check the correlation to calculate the current value of parameter (initial value plus amendment), and check whether the value has exceeds the reasonable range. If the value has already exceeded the reasonable range, the tuning fails.
Ensure no K parameters can exceed the range by experience listed in 3.2.4. Table 2.1 Value range of K parameters Tuning factor
Minimum
Maximum
Typical
K2
20
70
44.9
K3
–20
20
5.83
K4
0
1
0.5
K5
–10
0
–6.55
K6
–1
0
0
If the K parameters exceed the reasonable range, it is recommended to delete the tuned model and re-duplicate a model based on the configured model to be tuned. Start tuning with the configured model to be tuned. It is not recommended that you continue tuning by adjusting the K parameter which has exceeded the reasonable range. Every tuning starts with the default values of SPM. If the K parameter is within reasonable range after tuning, you can continue tuning by clicking the Identify button. When the K parameter is within the range by experience, the standard deviation is smaller than 8 by experience and the statistics result is stable, you can finish tuning. If the result is unsatisfactory, restart tuning with default values by adjusting parameters like effective height of antennas.
3. Auto Model Tuning After the automatic model tuning is complete, if the parameters fail to comply with the requirements in 3.2.4, you need to restart the model tuning in the manual mode. The process of automatic model tuning is as below: Step 1 Right-click the Explorer/Modules/ Models for Calibration/Calibration… in the U-Net, and select Automatic Calibration, as shown in 3.2.4.
2008-12-10
All rights reserved
Page50 , Total85
INTERNAL
Figure 1.1 Automatic Calibration
Step 2 Select the CW test data and parameters for calibration in turn, and click Commit to complete the setting, as shown in 3.2.4. Figure 1.1 Select parameters for calibration
3.2.5 Proposals on SPM Tuning Though each multiplier of SPM can be tuned, you cannot tune all the coefficients correctly at the 2008-12-10
All rights reserved
Page51 , Total85
INTERNAL
current stage due to limited collected data. The K1 and K2 are mandatory for model tuning, and whether to tune other parameters depends on the following proposals.
4. Coefficients Relevant to Effective Height of Antenna (K3/K5/K6) The K3 is relevant to effective height of antenna. The antenna height keeps fixed in measurement, the distance to the antenna is within 3 km, and the terrain changes a little, so the effective height of antenna changes little. Therefore, tuning K3 is not recommended. Similar with K3, tuning K5 is also not recommended. The K6 is relevant with the effective height of UE. The UE serves as a receiver in test, so K6 equals to 0 and its impact can be neglected.
5. Diffraction Multiplier (K4) The K4 is relevant to diffraction calculation. If the used map lacks of height information of buildings, the diffraction loss will be calculated based on the ground height. If you test within a small range, the terrain undulates a little. This differs greatly from knife-edge objects in calculating diffraction loss, so the calculation is inaccurate. As a result, tuning K4 is not recommended. −
If you test within a large range, the terrain undulates greatly. For example, the area is mountainous. As a result, tuning K4 is recommended. −
If you use high-resolution 3D maps with the height information about buildings, you can calculate diffraction loss with the height information. Therefore, tuning K4 is recommended. In this way, the obtained model will be more accurate. −
6. Kclutter and Clutter Loss The CW measurement proceeds in outdoor open land, but the points will be inadequate in other clutters. As a result, do not tune Kclutter and losses per clutter loss. Therefore, you can configure Kclutter to 1 and losses per clutter loss to 0, or Kclutter to 0. This has no impact on tuning result. In simulation, you need configure Kclutter to 1; you need configure losses per clutter loss according to conditions of digital maps or local conditions. Different digital maps contain the different clutters with different definitions, so you shall set them accordingly. In a planning project, the values of clutter losses must be confirmed by the operator or even provided by the operator. 3.2.5 lists typical values of clutter losses. Table 1.1 Typical values of clutter losses Clutter Type
Losses per clutter loss
Open Land in Village
0
Open Land in Urban
0
Wet Land
0
Village
6
Town in Suburban
8
2008-12-10
All rights reserved
Page52 , Total85
INTERNAL
Park in Urban
0
Parallel and Lower Buildings
18
Others Lower Buildings
13
Ocean Area
–2
Large and Lower Buildings
18
Inland Water
–2
High Buildings
20
Green land
0
Forest
10
Dense Urban
16
Common Buildings
13
7. LOS/NLOS The proposals about LOS/NLOS are similar with these for 3.2.5K4. If the used map lacks of height information about buildings and the terrain undulates a little in the calculated range, the model cannot distinguish LOS and NLOS. Therefore, distinguishing LOS/NLOS is not recommended. −
If you test within a large range, the terrain undulates greatly. For example, the area is mountainous. As a result, you can tune the model with LOS and NLOS respectively with two sets of parameters. −
If you use high-resolution 3D maps with height information about buildings, you can tune the model with LOS and NLOS respectively with two sets of parameters. −
For the later two cases, there must be enough spots (>200) for LOS and NLOS to guarantee accurate tuning. Before tuning, collect statistics of spots for LOS and NLOS respectively (by clicking the statistics button on the tuning properties tab) and check whether the number of spots meets the requirement. If the spots for LOS or NLOS are inadequate, it is recommended not to distinguish LOS/NLOS.
8. Near/Far region SPM is a macro cell model applicable for a large range of cell radius, but engineers usually perform tests in a range of 3 km. As a result, it is not to distinguish near and far region.
9. Proposals on Tuning Result The proposals on tuning result are as below:
The correlation value can not necessarily be 0 in tuning.
The range of K parameters and the requirement of standard deviation less than 8 mentioned previously are by experience and not absolute. For some special clutter, no matter how you tune parameters, the tuning result still fails to meet the reasonable range of K parameters and the requirement of standard deviation less than 8. A little deviation is also acceptable.
2008-12-10
All rights reserved
Page53 , Total85
INTERNAL
Model tuning is iterated process, sometimes it takes a long time. In addition, you do not necessarily tune K parameters according to correlation or big to small. When you fail to tune models ideally, you can try comprehensive tuning.
3.2.6 Analyzing Result and Verifying Model 10. Evaluating Model Deviation The main indexes of model tuning result are average error, standard deviation, and relative coefficients. The model tuning or verification shall meet the following conditions: −
−
Average error E : − 2dB ≤ E ≤ 2dB
Standard deviation: σE : σ ≤ 8dB
11. Analyzing Model Tuning Result After tuning, you shall analyze the tuning result, verify its validity, and evaluate its accuracy. There are two methods.
Checking with Statistics Report: SPM supports outputting model statistics report for analysis, and this report includes the information like standard deviation.
Illustrating model tuning: By illustrating, you can clearly see the distribution of model deviation so that you can evaluate the validity of model. For example, in the diagram, the deviation on the roads is large, so you can judge that the large deviation is due to test errors. You can delete these points and re-tune the model.
1) Checking with Statistics Report After tuning, in the calibration window, click Statistics button. A Report window is displayed, as shown in 3.2.6. The statistics report covers model parameters, clutter losses, total deviation of model, and deviation of various clutters.
2008-12-10
All rights reserved
Page54 , Total85
INTERNAL
Figure 1.1 Statistics report after tuning
2) Comparison with Correlation Curve Right lick the measurement data of tuned model, and a column is displayed, as shown in 3.2.6..
2008-12-10
All rights reserved
Page55 , Total85
INTERNAL
Figure 1.2 Measurement parameters of tuned model
A drop-down list is displayed with the following actions:
Calculate Signal Levels: calculate the level of signal from the transmitter according to the selected model. After calculation, U-Net automatically collects overall statistics and statistics of clutters, and then displays them.
Refresh geo data: refresh map information, such as clutter type and height.
Display statistics: display statistics of calculate prediction. It does not calculate prediction again.
Engineers usually use the two actions as below:
Calculate Signal Levels (mandatory)
Display statistics
Detailed operations of Calculate Signal Levels are as below: Step 1 Select the tuned model in the Propagation model drop-down list, as shown in 3.2.6. Step 2 Click the Calculations button. Step 3 Select Calculate Signal Levels. Step 4 Right-click the measurement data after calculation.
2008-12-10
All rights reserved
Page56 , Total85
INTERNAL
Step 5 Select Open the Analysis Tool, and a comparison curve is displayed, as shown in 3.2.6. Figure 1.1 Comparison curve
The red curve stands for the measured value. The blue curve stands for the predicted value of tuned model. You can see the distribution of model deviation with the suitability of two curves.
3) Comparison with Error Distribution Chart Right-click the corresponding measurement data of tuned model, and the properties window is displayed. Click the Display tab, as shown in 3.2.6. Figure 1.2 Properties window of measurement data of tuned model
Select the display type and field as shown in 3.2.6. Error(P-M) stands for the error between predicted value (P) and measured value (M). You can adjust the error range highlighted by different colors in 3.2.6; you can see the error distribution, as shown in 3.2.6.
2008-12-10
All rights reserved
Page57 , Total85
INTERNAL
Figure 1.3 Error distribution
You can also predict coverage with the model and see the distribution and variation of predicted field. They indicate the result of model tuning. The details are neglected here.
12. Verifying Model Verifying model is applying tuned model to an unknown area, comparing the predicted value and measured value, and obtaining an error evaluation (section 3.2.6). The requirements on average error, standard deviation, and relevant coefficients described in 3.2.6 shall be met. In actual operations, when selecting CW test sites, select another site for the verification spot for a model. The site shall meet all the features and conditions for model tuning spot. The CW measurement data of the site does not serve model tuning, but the method to process the CW measurement data of the site is the same as that for other sites. The data for verifying the site and the data of tuned site are imported together. You can adjust the detailed filtering conditions accordingly. In the Assisted Calibration window, the standard deviation between predicted value and measured value was shown in the lower left corner. You can also obtain the standard deviation from statistics result of statistics, as well as correlation curve (refers to the section Comparison with Correlation Curve) and error distribution (Comparison with Error Distribution Chart) refers to the section.
2008-12-10
All rights reserved
Page58 , Total85
INTERNAL
4
Volcano Model Tuning
4.1 Configuring Parameters of Volcano Model Volcano models, ray tracing models, are developed by Siradel. Volcano models use hybrid method of LOS prediction and ray tracing. They calculate path loss with enhanced Deygout method for some multi-path of LOS and ray tracing. They introduce tuning coefficients based on experience model and can be tuned with CW measurement data. They support maps of different resolutions in different areas. Volcano scenarios include the following three types:
Macro cell: the Tx antenna is higher than surrounding buildings.
Micro cell: the Tx antenna is lower than surrounding buildings.
Mini cell: the Tx antenna is between the Tx antenna of Macro cell and that of Micro cell.
For the second and third scenarios, Volcano models use ray tracing for calculation. For the macro cell scenario, the Volcano model, similar with SPM, calculates LOS loss only. You can install Volcano respectively. After installation, the Propagation Models list in the Modules tab contains three models:
Volcano Macrocell
Volcano Microcell
Volcano Minicell
2008-12-10
All rights reserved
Page59 , Total85
INTERNAL
Figure 1.1 Volcano models displayed in U-Net
4.1.1 Configuring Parameters of Volcano Macrocell Model In U-Net, in the Explorer window, in the Modules tab, in the unfolded Propagation Models list, right-click the Volcao Macrocell model. Select Properties. You can configure its properties in the macro properties window.
13. Parameters in General Tab You can configure the name and description information of the model, similar to these of SPM.
2008-12-10
All rights reserved
Page60 , Total85
INTERNAL
Figure 1.1 Parameters in General Tab for Volcano Macrocell model
11
Parameters in Map Data Tab
You can configure the parameters related to map used in model and map-related aspects, as shown in 4.1.1.
2008-12-10
All rights reserved
Page61 , Total85
INTERNAL
Figure 1.1 Parameters in Map data tab for Volcano Macrocell model
Vertical analysis mode: it indicates whether the height information about clutters is from 3D raster map (raster favorite) or 3D vector map (vector favorite). If you use 2D vector map without height information about clutters, you shall select raster favorite. Map data layers: the digital map used by the model. You shall import layers into Volcano model respectively. The layers are usually consistent with the imported digital maps in U-Net. Altimetry: raster map of terrain height. In digital maps, it is contained in the index file under the heights or DTM directory. Clutter: raster map of clutter type. In digital maps, it is contained in the index file under the clutter or DLU directory. Clutter height: raster map of clutter height. In digital maps, it is contained in the index file under the building or DHM directory. 3D Vector: 3D vector map. In digital maps, it is contained in the index file under the vector directory. The vector map in Volcano Macrocell model is optional. In vertical analysis mode, when selecting vector favorite, you shall import 3D vector map, because the model requires abstracting height information of clutters from 3D vector map. Vector reference: it is valid only when 3D vector map is imported. It indicates whether the height reference of vector map is ground (relative height) or sea surface (absolute height). Its reference is usually ground. Prediction Preferences: it is valid only when the import map contains various resolutions, such as 20 m and 5 m. It indicates the preferential resolution.
2008-12-10
All rights reserved
Page62 , Total85
INTERNAL
12
Parameters in Clutters Tab
You can configure clutter parameters in raster map in Clutter tab, as shown in 4.1.1. Figure 1.1 Parameters in Clutters Tab for Volcano Macrocell model
When you import the raster map of clutter type (DLU), you can see various clutters in the Clutter tab. Volcano Macrocell model describes raster maps from the following aspects:
Volcano type: the clutter type defined in Volcano. There are several clutter types. They are Land, Water, Building, Vegetation, Bridge, Built-up area. These six clutter types use different calculation strategies. You shall select corresponding volcano type according to the definitions of clutter types in digital maps.
Clutter attenuation: the clutter loss. Each volcano type has default clutter attenuation, but you can change it.
Clutter height: the clutter height. If there is no 3D map, you can specify a uniform clutter height for each clutter type. It is invalid when you use 3D maps.
2008-12-10
All rights reserved
Page63 , Total85
INTERNAL
13
Parameters in Vectors Tab
You can configure vector parameters of vector maps in the Vectors tab, as shown in 4.1.1. If you use 2D vector map or no vector map, this tab is invalid. Figure 1.1 Parameters in vectors tab for Volcano Macrocell model
When importing a vector map, you can see various clutter types in the Vectors tab. Volcano Macrocell model describes vector properties from the following aspects:
Volcano type: the vector type defined in Volcano model. The Volcano types are Land, Water, Building, Vegetation and Bridge. These five clutter types use different calculation strategies. You shall select corresponding volcano type according to the definitions of vector types in digital maps.
2008-12-10
Clutter attenuation: the clutter loss. Each volcano type has a default clutter loss, but you can change it.
All rights reserved
Page64 , Total85
INTERNAL
14
Parameters in Parameters Tab
You can configure algorithm parameters in Parameter tab for Volcano Macrocell model. Figure 1.1 Parameters in Parameter tab for Volcano Macrocell model
Free space correction: coefficient for free space correction. You can set two sets of A and B for LOS and NLOS respectively. Deterministic weighting: weighting factor of deterministic calculation. Environment: environment tuning parameter. If you use low-resolution map without clutter height layer (20-meter resolution map), you can select Low resolution; otherwise, you select Urban. Geographic profile extraction: the algorithm to extract lateral section with Deygout method.
If you select radial, you will abstract the lateral sections between the transmitter and the center of all rasters. For any receiver spots, select the nearest lateral section. Engineers usually select radial.
If you select systematic, you will abstract the lateral section between the transmitter and
2008-12-10
All rights reserved
Page65 , Total85
INTERNAL
receiver for all receivers. Therefore the calculation amount is great. K factor: the amendment factor of earth curvature towards effective height of antenna. You can use the default value. Indoor penetration: indoor prediction. The "indoor" referred herein is inside buildings of building type.
15
Parameters in Tuning Tab
You can configure model tuning parameters in Tuning tab and tune model, as shown in 4.1.1. Figure 1.1 Parameters in Tuning tab for Volcano Macrocell model
Autotuning mode: it indicates simple tuning or full tuning. Simple tuning are for the free space correction A and B, and deterministic weighting alpha. Full tuning is for all parameters include clutter loss and clutter height. Statistical tuning: It is valid when you select full tuning. It indicates whether to tune clutter loss and clutter height. Huawei performs CW measurements in outdoor open land, so selecting No in both
2008-12-10
All rights reserved
Page66 , Total85
INTERNAL
Attenuation and Height drop-down lists is recommended. Tune parameters: you can start model tuning by clicking Tune Parameters button.
4.1.1 Configure Parameters of Volcano Microcell Model In U-Net, right-click the Volcao Microcell model in the Explorer\Modules\Propagation Models list. Select Properties. You can configure its properties in the micro properties window.
16
Parameters in General Tab
You can configure the name and description information of the model, similar to these of SPM. Figure 1.1 Parameters in General Tab for Volcano Microcell model
17
Parameters in Map Data Tab
You can configure the parameters related to map used in model and map-related aspects, as shown in 4.1.1.
2008-12-10
All rights reserved
Page67 , Total85
INTERNAL
Figure 1.1 Parameters in Map data tab for Volcano Microcell model
Vertical analysis mode: it indicates whether the height information about clutters is from 3D raster map (raster favorite) or 3D vector map (vector favorite). If you use 2D vector map without height information about clutters, you shall select raster favorite. Map data layers: the digital map used by the model. You shall import layers into Volcano model respectively. The layers are usually consistent with the imported digital maps in U-Net. Altimetry: raster map of terrain height. In digital maps, it is contained in the index file under the heights or DTM directory. Clutter: raster map of clutter type. In digital maps, it is contained in the index file under the clutter or DLU directory. Clutter height: raster map of clutter height. In digital maps, it is contained in the index file under the building or DHM directory. 2DVector/3D Vector: 2D/3D vector map. In digital maps, it is contained in the index file under the vector directory. The vector map in Volcano Microcell model is mandatory, and at least the 2D vector map is mandatory. Vector reference: it is valid only when 3D vector map is imported. It indicates whether the height reference of vector map is ground (relative height) or sea surface (absolute height). Its reference is usually ground. Prediction Preferences: it is valid only when the import map contains various resolutions, such as 20 m and 5 m. It indicates the preferential resolution.
2008-12-10
All rights reserved
Page68 , Total85
INTERNAL
18
Parameters in Clutters Tab
You can configure clutter parameters in raster map in Clutter tab, as shown in 4.1.1. Figure 1.1 Parameters in Clutters Tab for Volcano Microcell model
When you import the raster map of clutter type (DLU), you can see various clutters in the Clutter tab. Volcano Microcell model describes raster maps from the following aspects:
Volcano type: the clutter type defined in Volcano. The clutter types are Land, Water, Building, Vegetation, Bridge and Built-up area. These six clutter types use different calculation strategies. You shall select corresponding volcano type according to the definitions of clutter types in digital maps.
Clutter attenuation: the clutter loss. Each volcano type has a default clutter attenuation, but you can change it.
Building Linear Loss: linear loss of buildings. You can configure it for the clutter of building type. It is 0.5 dB/m by default. If you do not consider linear loss of buildings, you can configure it to 0.
In Volcano Microcell model, the penetration loss of buildings includes two parts: clutter attenuation and building linear loss, but in SPM, there is loss per clutter class only. To make Volcano model and SPM compatible, you can configure the clutter attenuation of Volcano consistent with losses per clutter class of SPM while you configure building linear loss to 0. In this way, SPM and Volcano
2008-12-10
All rights reserved
Page69 , Total85
INTERNAL
model have same indoor penetration loss.
19
Parameters in Vectors Tab
You can configure vector parameters of vector maps in the Vectors tab, as shown in 4.1.1. If you use 2D vector map or no vector map, this tab is invalid. Figure 1.1 Parameters in vectors tab for Volcano Microcell model
When importing a vector map, you can see various clutter types in the Vectors tab. Volcano Macrocell model describes vector properties from the following aspects:
Volcano type: the vector type defined in Volcano model. The Volcano types are Land, Water, Building, Vegetation and Bridge. These five clutter types use different calculation strategies. You shall select corresponding volcano type according to the definitions of vector types in digital maps.
2008-12-10
Clutter attenuation: the clutter loss. Each volcano type has a default clutter loss, but you can change it.
All rights reserved
Page70 , Total85
INTERNAL
You must configure volcano type to building in vector maps; otherwise, Volcano Microcell model will not perform multipath calculation of ray tracing.
If some 2D vector map with 5-meter resolution, the Vectors tab is invalid, gray. For the vector of building type, you shall check whether its volcano type is Building. 4.1.1 shows the correct configuration. Figure 1.2 Correct configuration of vectors of building type
If the Volcano type (especially the building type) is incorrect in the Vectors tab, you need modify the menu file for vector map by adding #BUILDING at the building type, as shown in 4.1.1. After modification, you need re-import the vector layer.
2008-12-10
All rights reserved
Page71 , Total85
INTERNAL
Figure 1.3 Menu file for vector map
20
Parameter in Ray Tracing Tab
You can configure parameters of ray tracing algorithm for Volcano Microcell model, as shown in 4.1.1. Figure 1.1 Parameters in Ray Tracing tab for Volcano Microcell model
Launching parameters: the parameters of ray tracing (transmission ray) algorithm used by the model.
2008-12-10
All rights reserved
Page72 , Total85
INTERNAL
Maximum number of diffractions: maximum number of diffractions usually configured to 1. Maximum number of reflections: maximum number of reflections usually configured to 4. Angular step: angular step of transmission ray, usually configured to 1.5°. Ray stopping limits: maximum distance of ray propagation.
21
Parameters in Parameters Tab
You can configure other parameters for Microcell model, as show in 4.1.1. Figure 1.1 Parameters in Parameter tab for Volcano Macrocell model
Free space correction: coefficient for free space correction. You can set two sets of A and B for LOS and NLOS respectively. Deterministic weighting: weighting factor of deterministic calculation. Urban correction: environment correction factor. It includes forward correction, backward correction, and roughness.
2008-12-10
All rights reserved
Page73 , Total85
INTERNAL
You can obtain the values of previous parameters by model tuning. Geographic profile extraction: the algorithm to extract lateral section with Deygout method.
If you select radial, you will abstract the lateral sections between the transmitter and the center of all rasters. For any receiver spots, select the nearest lateral section. Engineers usually select radial.
If you select systematic, you will abstract the lateral section between the transmitter and receiver for all receivers. Therefore the calculation amount is great.
Indoor penetration: indoor prediction. The "indoor" referred herein is inside buildings of building type. Tune parameters: you can start model tuning by clicking Tune Parameters button.
4.1.2 Configuring Parameters of Volcano Minicell Model In U-Net, in the Explorer window, in the Modules tab, in the unfolded Propagation Models list, right-click the Volcao Minicell model. Select Properties. You can configure its properties in the mini properties window.
22
Parameters in General Tab
You can configure the name and description information of the model, similar to these of SPM.
2008-12-10
All rights reserved
Page74 , Total85
INTERNAL
Figure 1.1 Parameters in General Tab for Volcano Miniocell model
23
Parameters in Map Data Tab
You can configure the parameters related to map used in model and map-related aspects, as shown in 4.1.2.
2008-12-10
All rights reserved
Page75 , Total85
INTERNAL
Figure 1.1 Parameters in Map data tab for Volcano Microcell model
Vertical analysis mode: it indicates whether the height information about clutters is from 3D raster map (raster favorite) or 3D vector map (vector favorite). If you use 2D vector map without height information about clutters, you shall select raster favorite. Map data layers: the digital map used by the model. You shall import layers into Volcano model respectively. The layers are usually consistent with the imported digital maps in U-Net. Altimetry: raster map of terrain height. In digital maps, it is contained in the index file under the heights or DTM directory. Clutter: raster map of clutter type. In digital maps, it is contained in the index file under the clutter or DLU directory. Clutter height: raster map of clutter height. In digital maps, it is contained in the index file under the building or DHM directory. 2DVector/3D Vector: 2D/3D vector map. In digital maps, it is contained in the index file under the vector directory. The vector map in Volcano Microcell model is mandatory, and at least the 2D vector map is mandatory. Vector reference: it is valid only when 3D vector map is imported. It indicates whether the height reference of vector map is ground (relative height) or sea surface (absolute height). Its reference is usually ground. Prediction Preferences: it is valid only when the import map contains various resolutions, such as 20 m and 5 m. It indicates the preferential resolution.
2008-12-10
All rights reserved
Page76 , Total85
INTERNAL
24
Parameters in Clutters Tab
You can configure clutter parameters in raster map in Clutter tab, as shown in 4.1.2. Figure 1.1 Parameters in Clutters Tab for Volcano Minicell model
When you import the raster map of clutter type (DLU), you can see various clutters in the Clutter tab. Volcano Minicell model describes raster maps from the following aspects:
Volcano type: the clutter type defined in Volcano. The clutter types are Land, Water, Building, Vegetation, Bridge and Built-up area. These six clutter types use different calculation strategies. You shall select corresponding volcano type according to the definitions of clutter types in digital maps.
Clutter attenuation: the clutter loss. Each volcano type has a default clutter attenuation, but you can change it.
Building Linear Loss: linear loss of buildings. You can configure it for the clutter of building type. It is 0.5 dB/m by default. If you do not consider linear loss of buildings, you can configure it to 0.
In Volcano Minicell model, the penetration loss of buildings includes two parts: clutter attenuation and building linear loss, but in SPM, there is loss per clutter class only. To make Volcano model and SPM compatible, you can configure the clutter attenuation of Volcano consistent with losses per clutter class of SPM while you configure building linear loss to 0. In this way, SPM and Volcano model have same indoor penetration loss.
2008-12-10
All rights reserved
Page77 , Total85
INTERNAL
25
Parameters in Vectors Tab
You can configure vector parameters of vector maps in the Vectors tab, as shown in 4.1.2. If you use 2D vector map or no vector map, this tab is invalid. Figure 1.1 Parameters in vectors tab for Volcano Minicell model
When importing a vector map, you can see various clutter types in the Vectors tab. Volcano Macrocell model describes vector properties from the following aspects:
Volcano type: the vector type defined in Volcano model. The Volcano types are Land, Water, Building, Vegetation and Bridge. These five clutter types use different calculation strategies. You shall select corresponding volcano type according to the definitions of vector types in digital maps.
Clutter attenuation: the clutter loss. Each volcano type has a default clutter loss, but you can change it.
Building Linear Loss: linear loss of buildings You can configure it for the clutter of building type. It is 0.5 dB/m by default. If you do not consider linear loss of buildings, you can configure it to 0.
2008-12-10
All rights reserved
Page78 , Total85
INTERNAL
You must configure volcano type to building in vector maps; otherwise, Volcano Microcell model will not perform multipath calculation of ray tracing.
If some 2D vector map with 5-meter resolution, the Vectors tab is invalid, gray. For the vector of building type, you shall check whether its volcano type is Building. 4.1.2 shows the correct configuration. Figure 1.2 Correct configuration of vectors of building type
If the Volcano type (especially the building type) is incorrect in the Vectors tab, you need modify the menu file for vector map by adding #BUILDING at the building type, as shown in 4.1.2. After modification, you need re-import the vector layer.
2008-12-10
All rights reserved
Page79 , Total85
INTERNAL
Figure 1.3 Menu file for vector map
26
Parameter in Ray Tracing Tab
You can configure parameters of ray tracing algorithm for Volcano Minicell model, as shown in 4.1.2. Figure 1.1 Parameters in Ray Tracing tab for Volcano Minicell model
Launching parameters: the parameters of ray tracing (transmission ray) algorithm used by the Minicell model.
2008-12-10
All rights reserved
Page80 , Total85
INTERNAL
Maximum number of diffractions: maximum number of diffractions, usually configured to 1. Maximum number of reflections: maximum number of reflections, 2 by default, usually configured to 4. Angular step: angular step of transmission ray, usually configured to 1°. Linear step: the distance between transmission rays, usually 4 m. Ray stopping limits/Width of near region: radius of near region.
27
Parameters in Parameters Tab
You can configure other parameters for Minicell model, as show in 4.1.2. Figure 1.1 Parameters in Parameter tab for Volcano Minicell model
Free space correction: coefficient for free space correction. You can set two sets of A and B for near region and far region respectively. Deterministic weighting: weighting factor of deterministic calculation. Ray-Tracing Weighting: weighting factor of ray tracing calculation. It includes reflection weight, diffraction weight, and backward weight.
2008-12-10
All rights reserved
Page81 , Total85
INTERNAL
You can obtain the values of previous parameters by model tuning. Indoor penetration: indoor prediction. The "indoor" referred herein is inside buildings of building type. Tune parameters: you can start model tuning by clicking Tune Parameters button.
4.2 Tuning Volcano Models The flow for tuning Volcano models in U-Net is similar with that for tuning SPM. The flow also includes the following steps:
28 29 30 31 32
Set up a tuning project.
Import map.
Set sites and antennas
Import and adjust CW measurement data.
Set up propagation model
----End The detailed steps are skipped herein, which are in the sections 3.2and 3.2.3. This section describes the process of model tuning and the method to check tuning result.
2008-12-10
All rights reserved
Page82 , Total85
INTERNAL
4.2.1 Tuning Process Before tuning Volcano model, you shall configure the initial values of model parameters, detailed in 4.1. Configure the parameters of Volcano models to the default values. CW measurements collect outdoor data, so you can skip indoor prediction to save tuning time by clearing indoor penetration check box in the parameters tab.
33
Volcano Tuning Dialog Box
After filtering data and setting model correctly, you can tune Volcano models. In the propagation model properties dialog box, click Tuning Parameters button, and a Volcano Tuning dialog box is displayed. The Volcano Macro/Micro/Mini model tuning dialog boxes are similar, as shown in 4.2.1. Figure 1.1 Volcano tuning dialog box
The only difference between micro model and other two models is the Autotuning mode drop-down list, which includes full tuning and Simple tuning.
2008-12-10
All rights reserved
Page83 , Total85
INTERNAL
Figure 1.2 Selecting automatic tuning mode for Volcano Microcell model
The automatic tuning mode of Volcano Macrocell model is selected in the Parameters tab. For Macrocell/Microcell model, the full tuning mode is recommended. Full tuning and simple tuning make no difference for Minicell model.
34
Automatic Model Tuning
As shown in 4.2.1, you can see the imported CW measurement data in the upper part of the dialog box. You can select one or more groups of data as required. After selection, start automatic tuning by clicking the Start button. The tuning time depends on the type of Volcano model (Macro/Micro/Mini) and amount of data. Generally, tuning Macrocell model takes least time, usually within half a minute. Tuning Microcell/Minicell is slower, about 4-10 minutes.
4.2.1 Checking and Analyzing Tuning Result After model tuning, a dialog box as shown in 4.2.1 pops up. Figure 1.1 Tuning report
The Tuning Report dialog box displays the tuned parameters. The most important parameters are A
2008-12-10
All rights reserved
Page84 , Total85
INTERNAL
and B. The range of tuned A is [–5, 5] while that of tuned B is [20, 30]. Apply the tuned parameters to the model by clicking Apply button. You can see open a detailed tuning report by clicking Show results, and the report is a text file. Figure 1.2 Calibration result
The text file includes the following aspects:
Standard deviation
Correlation factor
Clutter result
CW measurement proceeds in open land, so there must be an error due to location error if a point is on other clutters. If these points have great impact on result, you shall re-tune after deleting these points. The displayed diagram of tuning result and steps to verify result are similar to these of SPM. You can refer to 3.2.6.
2008-12-10
All rights reserved
Page85 , Total85
