Home

User Manual

Contents

1. name Lat Lon proj rL datum wgs84 ellps wgs84 no_defs defined 31 PROJ_UNITS unit degree units degrees meters Ss 10 GRASS 6 4 0 LongLat GRASS GIS gt Radio coverage calculations depend on distances in meters and hence require a location with a cartographic projection SRTM maps on the other hand represent small one angular degree patches using unprojected coordinates in angular degrees Hence a Longitude Latitude GRASS location must be created to handle and process these maps After a GRASS DEM map is created from a set of SRT maps it can be projected to the final GRASS location with cartographic projection using the GRASS r proj command All projections of coordinates and whole GRASS raster maps between two coordinate systems are performed by the PROJ 4 library and the GDAL part of the related GDAL OGR library In the next subchapter we will explain the tools and procedures for e downloading the necessary SRTM maps e joining them into a GRASS map in a Latitude Longitude pseudo projection GRASS location e projecting the GRASS map to the final cartographic location e optionally removing void regions by interpolation e creating a dummy clutter map in the absence of a real clutter map required for r haraDEM At the end an example is given in which a SRTM based DEM map is generated for Slovenia and compared with the of
2. 20 km e effective BS transmitter antenna height 30 200 m e effective MS receiver antenna height 1 10 m The model is based on the Hata model for the suburban areas L dB 46 33 33 9 log f MHz 13 82 log h m a h 44 9 6 55 log h m log d km C where C ZO for medium sized cities and suburban areas 3 for large cities centers h difference between the transmitter and receiver antenna heights h receiver antenna height above the ground d horizontal distance between the transmitter and the receiver f transmission frequency in MHz The height correction factor a h is given by a h 0 og f MHz 0 7 h m 1 56 log f MHz 0 8 The model is adjusted for higher transmission frequencies It is mostly suitable for medium sized and large cities and assumes the transmit base station antenna to be positioned above the surrounding buildings The model only partially takes into account the terrain configuration the effective height h in the equation 6 and ignores the signal loss behind large obstacles An example of a path loss map obtained with r cost23 is shown in Fig 5 14 Fig 5 Path loss at 2 GHz computed with r cost231 e Usage r cost231 q inputDEM name output name coordinate x y ant_height value radius value area_type string frequency value overwrite verbose quiet e Flags g Ouiet Allow output files to overwrite existing files Se Verb
3. 20log f MH2 1 where FSPL loss in dB R distance between the transmitter and the receiver f transmission frequency in MHz The model takes into account LOS Line of Sight 1 e the visibility between the transmitter and the receiver but in general this is a very simplified theoretical model that works fine in empty space but does not give accurate results in real terrestrial propagation environments where the signal loss deviates from the free space squared distance law it generally increases with distance with a higher exponent than 2 An example of a path loss map obtained with r fsp is shown in Fig 3 The transmitter is placed at the IJS location in Ljubljana computation is limited to 10 km around the transmitter the actual command used is listed in the Example below The same holds for the other models described in the next chapters 11 Fig 3 Path loss at 2 GHz computed with r fspl e Usage r fspl q inputDEM name output name coordinate x y ant _height value frequency value radius value overwrite verbose quiet e Flags q Quiet SO Allow output files to overwrite existing files y Verbose module output seg Quiet module output e Parameters inputDEM Name of input raster map output Name for output raster map coordinate Base station coordinates ant_height Height of the antenna m default 10 frequency Frequency MHz radius Radius of calculation km default 10 e Exa
4. ewres 100 rows 1700 cols 2600 cells 4420000 The official DEM is shown in Figs 12 and 13 with Fig 12 showing also the geographic map Google Map of the region for reference Fig 12 Commercial Slovenian DEM Google Map 35 e MG w z a ft KK n SZ de ee SC y ze ite D PR D s a a d 2836 2115 1394 673 A Ze 48 Au Fig 13 Official commercial Slovenian DEM Fig 14 shows DEM created from the SRTM maps for the same region and with the same color scale and color legend The only obvious differences are significant voids white areas mostly in the northwest mountainous region d Z n Ma Ac fe 5 3 2836 2115 1394 673 i ae 48 VU Fig 14 Slovenian DEM created from SRTM maps Fig 15 shows height differences between the official and SRTM based DEMs White areas represent the void areas of the SRTM based DEM They are accompanied with some small areas of large height errors reddish greenish colored reaching 436 to 366 m The prevalent color however is blue which corresponds to relatively small errors 36 Fig 15 Differences between the comercial and SRTM based Slovenian DEMs Figs 16 to 18 give some more insight into the SRTM based DEM height error magnitude Figs 16 and 17 show height errors along two diagonals from lower left corner to upper right corner Fig 16 and form upper left corner to lower right corner Fig 17 We can see that errors of
5. e Flags q Quiet 1 Rx dBm values in output map replaced by 1 0 when above rx_threshold o Allow output files to overwrite existing files PEN Verbose module output q Quiet module output e Parameters cell input Cells data table rx threshold inimum received power dBm for radio signal coverage default 999 output ame for output raster map table Table name driver Driver name options mysql ogr pg dbf sqlite none default none database Database name default SGISDBASE SLOCATION NAME SMAPSET dbf cell num umber of successive path loss values to be written in the table dbperf Database insert performance rows INSERT 99 special fast mode via CSV options 1 99 default 20 24 e Examples the first one does not create a data table but dummy values for the table and cell_num parameters must be specified anyway the second one creates a DBF data table named ijs_abc in the default GRASS dbf folder within the user s mapset r MaxPower cell input cell list output ijs_ abc table ijs_ abc cell num 5 o r MaxPower cell input cell list output ijs_ abc table ijs_ abc driver dbf dbperf 1 cell num 5 o 2 4 1 The input cell list file The cell list file Cells data table is specified by the cell input parameter It is a text file in a CSV like actually Semicolon Separated Values format where each line contains data for a transmission antenna here also called cell lt ce
6. file here the first line is split into three lines to fit on the page but is actually a single line Terrene type loss factors hataDEM model 1 irrigated agriculture 2 rangeland 3 coniferous forest 4 deciduous forest 5 mixed forest 6 disturbed Ltb 2 15 S25 4 20 27 Fig 10 shows the official commercial Slovenian land use map for the Ljubljana region it includes 12 categories The corresponding clutter map generated by r clutconvert is shown in Fig 11 D Jo E un wo a ER H Fig 11 The corresponding clutter map generated by r clutconvert e Usage r clutconvert input name Path loss values name output name overwrite verbose quiet e Flags 28 0 Allow output files to overwrite existing files v Verbose module output q Quiet module output e Parameters input Name of input raster map Path_loss values Path loss factors for land use output Name for output raster map e Example r clutconvert input clut category Path loss values home userl convtable output clut dBloss o 29 3 SRTM maps SRTM Shuttle Radar Topography Mission map are Digital Elevation Maps DEMs obtained by a special radar system onboard Space Shuttle Endeavour during its 11 day mission in February 2000 6 7 The maps are publicly available 7 Two alternate sources of SRTM based maps are 19 and 20 Another interesting project produc
7. 20 m and even twice that much are not uncommon 60 40 20 Raster values 20 Kei o 50000 100000 150000 200000 250000 300000 Distance metres Fig 16 Profile difference lower left corner to upper right corner diagonal Raster values o o 50000 100000 150000 200000 250000 300000 Distance metres Fig 17 Profile difference upper left corner to lower right corner diagonal Fig 18 contains a histogram of height difference in Fig 15 It shows that the mean height error is approximatelly O and that the majority of errors are below about 30 m 37 Py ty ty ty ty ty D N E Mm Co 27 12 19 55 X AXIS Cell Values in tens Y AXIS Number of cells in hundreds of thousands Fig 18 Height difference histogram This height error is somewhat larger that could be expected from the SRTM accuracy specifications A closer look reveals that high error values are largely related to steep slopes where height errors are caused not only by direct height measurements errors but also by horizontal position errors E g for a very steep terrain of 45 a horizontal position error of 20 m which corresponds to only 20 of the pixel size for a 100 m map raster could result in a vertical position error of up to 20 m depending on the direction of the horizontal position error relative to the slope direction The void areas in the SRTM based map in Fig 14 can be removed by interpolation with the GRASS GIS r fillnulls comma
8. constraints are different for LOS and NLOS cases see below In the LOS case transmitter receiver visibility the loss within the street canyon is defined as L dB 42 64 26 log d km 20 log f MHz d km gt 0 02 12 The first constant corresponds to the empty space loss at the distance of 20 m The transmitter antenna height must be at least 30 m and there should be no obstacles in the first Fresnel zone The signal loss is exponential with the distance the exponent value is 2 6 In the NLOS case no direct visibility between the transmitter and the receiver the model uses the following parameters e transmitter height 4 4 m to 50 m e receiver height h 1 m to 3 m e buildings height br 3 m x number of floors plus 3 m for gabled roofs and O m for flat roofs e the transmitter antenna height above the roof height Ah h W8 e the receiver antenna height below the roof height Ah W8 h e spacing between buildings b if no data is available the recommended value is between 20 m and 50 m e street width w if no data is available the recommended value is b 2 e incident angle of radio rays if no data is available the recommended value is 90 The path loss is 18 sl Lig Luas Ltr 2O 13 Los Las t Lansa SO It consists of three components e the free space loss Lo e the rooftop to street diffraction loss L e the multiple screen diffraction loss Linsg The free space loss is
9. deleted 2 1 5 Database support The r radcov script supports writing the output data table which is performed by r MaxPower The related command line parameters cellnum db_driver database out_table and dbperf are equivalent to those of r MaxPower and are described there By default if db_driver is not defined or is set to none no data table is created The r radcov script does not check for the actually installed and GRASS supported database management systems Instead it has a fixed list of them none dbf GRASS own built in database mysql MySQL and pg PostgreSQL Of course MySQL and PostgresSQL can only be used if they are installed and the GRASS support for them is also installed 2 1 6 Other parameters There are some more r radcov command line parameters and flags that will be described here briefly 10 The radius_ovr and model our parameters override per cell settings in the previously described cell list table While settings there are individual for each transmit antenna cell this two parameters set the radius and model with parameters globally i e for all antennas The model_ovr parameter expects a comma delimited list without spaces consisting of the model name and its parameters as described in Table 1 e g model_ovr hata urban The parameter rx_treshold is equivalent to that of the rMaxPower module and is used directly by that module When specified the received signal is ignored in those raster p
10. equation for open areas mk land use related signal loss at the receiver location in dB KDFR contribution of wedge diffraction in dB a parameter related to KDFR JDFR diffraction loss due to the Earth surface curvature in dB The Okumura Hata path loss as defined by this model is HOA dB A0 Al log d km A2 log Heff m A9 log d km log Hef m 3 2 log 11 75 Hmlm P 44 49 log f MHz 4 78 log f MHz where AO A3 model tuning parameter Heff difference between the transmitter and receiver antenna heights Hm receiver antenna height above the ground d horizontal distance between the transmitter and the receiver f transmission frequency in MHz We implemented the single wedge loss as 16 10 11 where h height of the wedge above the direct line between transmitter and receiver d da the distances of the mobile and base stations from the wedge Since r hataDEM was originally intended for calculation on smaller geographic areas cellular networks with transmitter receiver distances of up to 35 km it ignores the effect of the Earth surface curvature Additionally we fixed the value of parameter a to a 1 An example of a path loss map obtained with r hataDEM is shown in Fig 7 u LE 54 Ney Ze Fig 7 Path loss at 2 GHz computed with r hataDEM e Usage r hataDEM q inputDEM name clutter name output name AQ value Al value A2 valu A3 value coordinate x y ant
11. high 3 or 90 m at the equator and they contain void areas without valid measurements The radar measurements were performed with 30 m x 30 m spatial sampling with the following minimal required accuracies quoted at the 90 level 22 e lt 16m absolute vertical height accuracy e lt 10 m relative vertical height accuracy e lt 20 m absolute horizontal circular accuracy 30 The actual accuracy of the generated maps is discussed in 23 and other documents e g 24 25 Manually selecting and downloading individual one degree SRTM maps and combining them into a larger DEM would be cumbersome and error prone so we created scripts to automate these tasks The user only specifies the final region extents and the corresponding SRTM repository folder with files and the rest is done automatically 3 1 SRTM maps and GRASS locations projections Positions on the earth surface are defined globally in angular degrees relative to the equator and the prime meridian going through Greenwich Why these coordinates are natural the third dimension the height in meters usually above the sea level is a bit more complicated It depends on the Earth form which is not exactly a sphere but is modeled as a spheroid ellipsoid of revolution The current standard defining its parameters is WGS 84 WGS Word Geodetic System Local geographic maps are made by projecting a part of the earth surface to a plane using different cartographic projection
12. optionally used by GRASS were also installed This GRASS version is officially a Release Candidate version but marked as stable and recommended It has some bugs corrected which makes it a better choice compared to 6 4 2 e RaPlaT GRASS add on The installed versions of software components are new supported and recommended Unfortunately they or their combination exhibit some minor bugs which are supposed to be corrected in future versions The two most obvious are e When exiting GRASS GUI Graphic User Interface Le closing the GRAS GIS Layer Manager window a message is displayed multiple times in the GRASS Terminal text window python nnnn LIBDBUSMENU GLIB WARNING Trying to remove a child that doesn t believe we re its parent There are no other related problems and the message can safely be ignored It is a known bug caused indirectly by an external standard library used by GRASS 40 e A GUI mouse problem in GRASS in connection with the new Ubuntu graphical desktop GUI Unity The problem appears in certain situations e g when the user in the GRAS GIS Layer Manager window tries to add a map to the display and wants to use the slider to browse the maps It fails since pressing the left button to hold and move the slider unexpectedly cancels the operation a possible workaround is to first press the right button followed by ESC to close the resulting pop up menu and then the left button or to browse ma
13. service postgresql start Besides the original PostgreSQL user manual 28 short and useful quick start guides can be found in the web e g 29 44 5 New GRASS and RaPlaT installation RaPlat C modules are currently distributed as source code for Linux environment only and must be compiled before use Standard precompiled GRASS Linux packages do not include the development support needed for compilation of third party modules so GRASS must be installed from its source code distribution The required installation and configuration procedure are described in this chapter They were used to prepare the RaPlaT virtual machine and would be used for installation on a standalone physical machine with Ubuntu 12 04 pre installed However this description should be taken only as guidelines and your mileage may vary slightly due to the constantly changing software updates and versions Installation of GRASS RaPlaT and related software described in this chapter was done on a freshly installed and updated Ubuntu 12 04 1 desktop 32 bit operating system In the example below it is assumed that the username of the primary user account the one with administration rights i e sudo permissions is user 5 1 Install MySQL amp PostgreSQL First we install the database management systems shortly databases that we want to use with GRASS so that GRASS make can find them and build the GRASS binaries properly In our case these are MySQL and PostgreS
14. shorthand form of region rast e region n _ e _ 8 _ w _ res _ or region n _ e _ s _ Wes res _ with _ standing for numerical values sets the computation region by explicitly defining its extents and resolution instead of all five values any subset of them can be set with the rest of them retaining the existing values The last three ways of computation region definition i e with exception of current and dem can be combined e g the computation region can be defined with a raster map and then the resolution modified The mechanism is the same as in the case of the GRASS g region command because this command is actually used for region setting after replacing with and with so see g region help for more detailed information After the computation region is defined with the above procedure and before it is actually set and used r radcov performs some refinements e unifies north south and east west resolution by taking the latter east west for both directions e rounds the resolution to the integer value in m e rounds the region extents so that the values of pixel center coordinates are multiples of the resolution value and that this new region does exceed the extents of the original one 2 1 3 Parallel execution support GRASS modules are generally single thread processes that execute on a single processor core This is also true for RaPLaT modules However under certain limi
15. shown two differentiate between the Ubuntu shell and PostgreSQL client commands userl ubuntu psql grass grass gt Al grass gt dt grass gt SELECT FROM pg_user grass gt q The PostgreSQL configuration files are in the etc postgresql 9 I main for version 9 1 The default main database directory is var lib postgresql 9 1 main If necessary the PostgreSQL server can be restarted stopped or started with the following commands 52 sudo service postgresql restart sudo service postgresql stop sudo service postgresql start 53 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 References GRASS GIS Wikipedia http en wikipedia org wiki GRASS_GIS GRASS GIS home page http grass osgeo org GRASS RaPlaT The Radio Planning Tool for GRASS GIS system Home page http www e6 1js si index php en software grass raplat Andrej Hrovat Igor Ozimek Andrej Vilhar Tine Celcer Iztok Saje Tomaz Javornik Radio coverage calculations of terrestrial wireless networks using an open source GRASS system WSEAS Transactions on Communications 2010 vol 9 no 10 pp 646 657 Igor Ozimek Andrej Hrovat Andrej Vilhar Tine Celcer Iztok Saje Tomaz Javornik GRASS RaPlaT an open source tool for radio coverage calculations V Joint Workshop on Wireless Communications 1 2 March 2011 Paris France JNCW 2011 S 1 IE
16. the standard MSI text format 18 The r radcov script first executes the required path loss model modules for the given set of transmitters as specified in the cell list file continues with calling r sector for all the transmission antennas and finishes with calling r MaxPower for calculation of the overall radio signal coverage RaPLaT path loss model modules and r sector require a DEM map DEM Digital Elevation Map which describes the terrain relief Some path loss models 7 hataDEM from the above modules additionally need a clutter map which describes the signal loss due to the land use buildings forests lakes etc Fig 1 depicts an additional module r clutconvert which is used for creation of clutter maps describing land use dependent signal loss from general land use GRASS raster maps wee em ee me Me eM e ewe ew e ee ee ee ee ee r clutconvert rfspl rhata rhataDEM li r cost231 rsector oo sm o mm ee ao an ao a es mm e E as ee ER ER mm mm w we ew em zm ew e vm ew em ew zm ee ee ee ee ee ee Fig 1 GRASS RaPlaT block diagram GRASS modules generally work in the so called current region which defines the geographic region extents and resolution resampling of input maps if necessary is done automatically The r radcov script lets the user set the computation region independently of the current region a temporal current region is established for the execution of the c
17. the user creating them This can be done with the following command sudo find var local grassdata type d exec chmod gts Be aware that by setting a group ID grass in our case other that the primary user group userl grassuser members of this group grass may gain unwanted write access to the other users mapsets hence it is wise to review the group write permissions on directories and files within the database directory tree 5 5 Set up MySQL PostgreSQL for use with GRASS RaPlaT RaPLaT module r MaxPower called by r radcov optionally creates an output data table If we want to use MySQL or PostgreSQL to store it the following must be done first e a database or more of them must be created to store these tables e GRASS users must be granted access to this database The following examples show how this was done for the RaPLaT demo VM More information about MySQL and PostgreSQL usage can be found in 26 27 and 28 29 respectively 50 In the examples below the traditional command line used interface is used Various GUI based MySQL and PostgreSQL management tools can be used instead e g MySQL Workbench and pgAdmin III for MySQL and PostgreSQL respectively 5 5 1 Set up MySQL We can create a MySQL database with the following command mysqladmin u root create grass p The parameter u root requires execution with the MySQL root account to obtain administration rights and the user is asked for it
18. 0 ese eeeeeee ener 38 20 Differences between the comercial and the SRTM based voids filled Slovenian DEMS ee lee Ee A oa ee 39 Tables Table 1 An example of the cell list table eesecesecsseceseceseceseeeseeseneeeseeeaeessaecaaecsaecaeesaeenaeenes 4 Table 2 Description of the cell list table columns 200 0 lec eeceseceseeeeeeeeeeesaeeeseeesaecsaeceaecaecnaeenseeens 5 Table 3 An example of the antenna types table Be eee ceeeceseceseeeseeeseeesseeeseecseecsaecaecsaecnaeesaeenes 8 Table 4 Parameters and their values for the Walfisch Ikegami model eee eeeeeseeseeeneeereeeeeees 20 Table 5 Output data table format nres oreson e E E EE e E TAA 25 ili 1 GRASS RaPlaT overview GRASS GIS 1 2 shortly GRASS Geographic Resources Analysis Support System is a free Geographic Information System GIS software used for geospatial data management and analysis image processing graphics maps production spatial modeling and visualization It is available as prebuilt packets for various Linux distributions MS Windows and Mac OS X as well as in source code RaPlaT Radio Planning Tool for GRASS 3 4 5 is an add on for GRASS for radio signal coverage calculation It uses the GRASS support for geographic environment terrain relief and other GRASS functionalities displaying etc important for radio coverage computations and display RaPlaT comprises a set of C modules small programs written in C specifically for the GRASS envir
19. 00200200200200 32 3 3 CONVERT SRTM Mars TO A GRASS DEM M SRTMTOGRASS cosa oc cs san eedos 33 3 4 PROJECT GRASS DEM TO THE FINAL CARTOGRAPHIC LOCATION 4 4 4444 34 3 REMOVING VOIDS EE 34 3 6 WHAT ABOUT LAND USE CLUTTER MAPS etedegeteegeeee ere dee deg RRE 34 3 7 AN EXAMPLE SRTM BASED DEM FOR SLONENIA ea seca ae cea aa acedo 35 4 VIRTUAL MACHINE GRASS amp RAPLAT PREINSTALLED sssooo00000000000000000006 40 4 1 PRECONFIGURED UBUNTU USER ACCOUNTS cccsssssscecsessececssssseecessnssececsssaeeeesssneees 41 4 2 ABOUT ER EE E CN 41 43 ABOUT GRASS DATABASE EE 42 4 4 r ABOUTMYS Q EE 42 4 5 ABOUT POSTGRES EE 43 5 NEW GRASS AND RAPLAT INSTALLATION 0 ccsssscssssccssssccssssscssssccssccscssssceees 45 5 1 INSTALL MYSOL amp Re ER EEN 45 5 2 INSTALL GRASS AND RELATED SOFTWARE PACKAGES B B 4 4 4 24 4 4 42242200 45 5 3 INSTAL RAPLA MODULES slacker BEN eh Sane Recht ee eile ie 48 5 4 SETUP GRASS DATABASE AND USERS Egeter Serebegre 49 5 5 SETUP MYSQL POSTGRESQL FOR USE WITH GRASS RAPLAT 4 4 44 50 IId Setup MySOLD ee Eeer lana da tega 51 55 27 Setup eegener 52 6 REFERENCES cen 54 eoccccce eoccccce eoccccce eoccccce 11 Figures Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig Fig 1 GRASS RaPlaT block diagram eee eee e
20. 10 WA Frequency correction 3 25 10 W5 Building height correction 10 30 20 W6 Width of roads m rec W7 2 15 W7 Building separation m rec 20 50 30 W8 Height of buildings m 12 The COST231 Walfish Ikegami provides good path loss estimates if the transmission antenna is located above the roof level If it is located near the ground level the estimates are bad because the model does not take into account the waveguide effect of the large city street canyons An example of a path loss map obtained with r waik is shown in Fig 8 de Co Fig 8 Path loss at 2 GHz computed with r waik e Usage r waik q overwrite verbose quite inputDEM name output name coordinate x y ant_height value frequency value radius value free space loss correction value bs _correction value range _correction value street width _correction value frequency correction value building height _correction value street _width value distance between buildings value building height value PHI_Street value area_type string e Flags 20 q Quiet 0 Allow output files to overwrite existing files v Verbose module output q Quiet module output e Parameters inputDEM Name of input raster map output Name for output raster map coordinate Base station coordinates ant_height Height of the antennas m default 10 frequency Frequency MHz radius Radius of calculat
21. 4 Pa PON STA POM NP OM PLO wp ya IJS A 17307 0 0 20 COS 21 4606097 99918 30 10 hata urban rrrrrriran TJS B 2 135 0 0 20 COS 21 460697 99918 30 10 hata urban 77 74 yy 4 TIS C 3 270 0 0 20 COS 21 460697 99918 30 10 hata urban rrrrrrri The first line contains the header in the example above it is split into three lines to fit on the page but should actually be a single line Each following line contains data for one transmission antenna The r radcov script parses this table according to a special data structure defined by a Python script variable cellTableDescrib which will not be explained here but is at least to a certain degree self explanatory The data columns their types and value constraints as defined by this structure are shown in Table 2 Table 2 Description of the cell list table columns Name Type Allowed values Description cellName name see description Cell name characters A Z a Z numbers _ antennalD id 1 999999 Antenna identification number beamDirection i 0 360 Antenna horizontal direction 0 northwards positive clockwise electricalTiltAngle i 0 10 Antenna electrical vertical tilt downwards mechanicalAntennaTilt i 90 90 Antenna vertical direction positive downwards heightAGL f 0 0 300 0 Antenna height above the terrain antennaType antype unconstrained Antenna type positionEas
22. 6 pc linux gnu To simplify calling of this script we created a symbolic link to it grass64 d d for development installation in usr local bin Hence GRASS is started simply by typing grass64 d in a Terminal window After starting GRASS creates an environment variable GISBASE which contains the complete path to dist i686 pc linux gnu This directory contains the whole GRASS binary distribution with everything needed to actually run GRASS This location within the GRASS source directory is suitable for development purposes including the development and compilation of additional modules like those from the RaPlaT source distribution The place for additional modules is in the doc subdirectory of the GRASS source directory Simple examples of such modules are already included there as part of the GRASS source distribution The second set of binaries would be created from dist i686 pc linux gnu make install in the normal system location usr lib grass64 and the GRASS start up script grass64 in usr bin This second set of binaries in the final system location is not really necessary to use GRASS and was not created in our RaPlaT VM 41 4 3 About GRASS database GRASS uses a DBF type database for storing its geographic maps and related information It has the form of a directory tree with the first level of subdirectories containing GRASS locations and the second level containing GRASS user s mapsets including the special PERMA
23. EE France section 2011 6 pp Shuttle Radar Topography Mission Wikipedia http en wikipedia org wiki Shuttle_Radar_Topography_Mission Shuttle Radar Topography Mission NASA Jet Propulsion Laboratory http www2 jpl nasa gov srtm VMware Player Home page http www vmware com products player Comma separated values Wikipedia http en wikipedia org wiki Comma separated_values Common Format and MIME Type for Comma Separated Values CSV Files RFC 4180 October 2005 http www rfc editor org rfc rfc4180 txt GRASS Wiki Category Parallelization http grasswiki osgeo org wiki Category Parallelization GRASS Wiki Parallel GRASS jobs http grasswiki osgeo org wiki Parallel GRASS_jobs S R Saunders Antennas and Propagation for Wireless communication systems M Hata Empirical formula for propagation loss in Land Mobile radio services IEEE Transactions on Vehicular Technology Vol 29 no 3 avgust 1980 D J Cichon T Kurner Propagation prediction models COST 231 Final Rep http www Ix it pt cost231 J Walfisch H L Bertoni A Theoretical Model of UHF Propagation in Urban Environments IEEE Trans Antennas Propagat Vol 36 pp 1788 1796 December 1988 J Walfisch H L Bertoni A Theoretical Model of UHF Propagation in Urban Environments IEEE Trans Antennas Propagat Vol 36 pp 1788 1796 December 1988 MSI Planet Antenna File Format http radiomobile pelmew nl The_program Definiti
24. GRASS and related software packages First update Ubuntu 12 04 with the latest updates Ubuntu Update Manager or commands sudo apt get update and sudo apt get upgrade Installation of GRASS from its source is performed according to 30 in the following steps 45 1 Install the required software packages dependencies libraries and other supporting software Use the following command it is a bit different from what is suggested in 30 for Ubuntu 12 04 sudo apt get install build essential make flex bison gcc libgccl g cmake ccache python python dev python qt4 python qt4 dev python opengl python wxversion python wxtools python wxgtk2 8 python dateutil libgsl0 dev python numpy wx2 8 headers wx common libwxgtk2 8 dev libwxgtk2 8 dbg libwxbase2 8 dev libwxbase2 8 dbg libncurses5 dev zliblg dev gettext libtiff dev libpngl2 dev tcl8 5 dev tk8 5 dev libcairo2 libcairo2 dev sqlite3 libsqlite3 dev libpq dev libreadline6 libreadline6 dev libfreetype6 dev txt2tags libfftw3 3 libfftw3 dev libqt4 core libqt4 dbg libqt4 dev libqt4 gui libqt4 sql libqt4 qt3support lsb qt4 qt4 designer qt4 dev tools qt4 doc qt4 qtconfig libapt pkg perl resolvconf libjasper dev ruby subversion ffmpeg ffimpeg2theora libffmpegthumbnailer dev libavcodec dev libxmu dev libavformat dev libswscale dev checkinstall libglul mesa dev libxmu dev And for the MySQL sup
25. HI Street Street orientation deg P11 s metropolitan Area type medium_cities Parameters P1 P11 for the fspl free space model P1 P11 not used Type means e name character string see description in the table Used for the cell names e id integer value similar to type i see below but values must be unique the same value may not repeat Used for antenna identification numbers in multi antenna systems cellular networks e antype character string allowed characters letters numbers and in any order Used to define for the antenna types e i integer value with min and max bounds If the value of both min and max bounds is 0 the value is unbounded The decimal point comma is not allowed ef floating point value with min and max bounds If the value of both min and max bounds is 0 0 the value is unbounded The value can be written in the cell list file without the decimal point comma e s a word character string form a set of allowed words The columns to the right can depend on this word as defined by the cellTableDescrib variable e g the Pn columns depend on the word model name in the model column e arbitrary contents Empty lines are ignored The character as the first character of the cell name or the first character in a line has a special meaning it marks the cell as a comment only effectively disabling the cell This is useful for simple
26. Institut Jozef Stefan Ljubljana Slovenija eo Department of Communication Systems E 6 4 GRASS RaPlaT Radio Planning Tool for GRASS User Manual V1 0a Igor Ozimek Andrej Hrovat Andrej Vilhar Tomaz Javornik Ljubljana September 2013 Contents 1 GRASS RA PLAT OVERVIEW sscsscsscsiscecscseosoesvensavseeevoncsssvuceveedovaqendvvvecessvseseoenseversoactensee 1 2 RAPLATIN DETAILS sisisiilados sedes cas didos dela aa da DAS DD onuetensselon coveos dedo be snzseessoucecevessescossvecccbenens 3 2 1 RUN A COMPLETE RADIO COVERAGE COMPUTATION R RADCOV 4 4 44444 4 De TEE 7 2 1 2 Computation E 8 213 Parallel execution EN 9 244 Reuse and Eege ee eege ans 10 239 Database SUupPOV svezi eet teen 10 2110 Other parameter S den al de ata il ne NG re aaah var rl RD 10 2 2 RADIO PROPAGATION MODELS ISOTROPIC ANTENNA cccccceeessessececececeeessensnseeeeeeees 11 E SE Il DDD EE 12 2 D3 ECO SED EE 14 PDA PA GIGDIEM EEN 15 DL E 18 2 3 ADD TRANSMISSION ANTENNA R SECTOR E 21 2 4 CALCULATE COMPLETE COVERAGE R MAXPOWER c ccccsssseceesesseeeeeesneeeeeeeeaaeeeees 24 DAL EE 25 24 2 EE 25 2 5 PREPARE CLUTTER MAP R CLUTCONVERT 1s iiu tidenivel iui Ve tia tenes beet ee 27 3 SRTM MAPS EE 30 3 1 SRTM MAPS AND GRASS LOCATIONS PROJECTIONS 1 4 4 4 22 4 4 2248222200200 31 3 2 DOWNLOAD SRTM MAPS M GETSRTMMAPS 4 2000000000
27. L W0 20 logd 20 log f 14 The rooftop to street diffraction loss is L 8 2 W3 log W6 W4 log f W5 log Ah L 15 where the orientation related loss is 10 0 354 0 lt lt 35 L 2 540 075 4 35 35 lt 655 16 4 5 0 11 55 55 lt lt 90 The multiple screen diffraction loss is La Loi k k logd k log f 9 oer WI 17 where the shadowing gain is 18 log1 Ah h gt h gf 3 t roof 18 0 h lt h roof The ka and ky parameters depend on the path length d and the transmitter height above the roofs Wi h 2 h o k 4W1 0 8 h h o h lt hop d 2 0 5m 19 W1 0 4 d h h h lt hp d lt 0 5m W2 E 15 h h ka w2 Au bh ER roof roof Parameter k represents the increase of path loss when the transmitter is located below the roof levels while parameters ky and ky represent the path loss due to distance and frequency The latter is defined by f k 4 k 1 21 f loss 21 The value of kg is 1 5 for city centers and 0 7 elsewhere Parameters W0 W8 should be set according to the recommended values given in Table 4 19 Table 4 Parameters and their values for the Walfisch Ikegami model Parameter Description Value range Default value WO Free space loss correction 20 60 32 5 Wi Reduced base antenna height correction 30 70 54 W2 Range correction 5 35 10 W3 Street width correction 3 15
28. NENT mapsets owned by the GRASS administrator A suitable system location for this database is in var local complete database path var local grassdata The VM database contains two demo locations e Slovenia with a cartographic projection suitable for this geographic region It contains mapsets PERMANENT containing a SRTM based DEM for Slovenia user and grassuser e LongLat with Longitude Latitude pseudo projection which can be used for processing SRTM maps as described earlier It contains only one mapset PERMANENT for use by userl If GRASS DBF database is used also for output data tables optionally created by RapPlaT they are normally stored in the dbf subdirectory within the current user s mapset Since each table is an independent DBF format file it could also be stored in another location e g the user s home directory 4 4 About MySQL The MySQL database in the RaPLaT VM already contains a pre created database named grass which can be used to store data tables created by RaPlaT Both preconfigured local users user and grassuser have full access grant all on this database The following commands can be used in a Terminal window to list the existing MySQL databases and data tables the pre created grass database in initially empty To login to MySQL and simultaneously connect the grass database mysql grass This is a shortcut for first performing login to MySQL and then connecting the grass datab
29. QL We can install both databases in the Ubuntu standard way with Ubuntu Software Center For MySQL search for mysql mysql server and mysql client for PostgreSQL search for postgresql During MySQL installation you are asked for the MySQL root user password Note that this has nothing to do with the Ubuntu root user account which is by default disabled for login has no password defined and sudo is used instead for administrative tasks Any Ubuntu user that knows the MySQL root password can gain administrative MySQL access PostgreSQL has a different approach to gain administrative access Its administrative permissions are assigned to a special Ubuntu user account postgres which is created during its installation This account is by default disabled for login and no password is defined but it can be used with sudo Hence no PostgreSQL password is required by default for the administrative access but it is available only to the users with Ubuntu administrative rights those with sudo permissions Originally MySQL and PostgreSQL have only command line interfaces If you prefer GUI based management you can additionally install MySQL Workbench search for mysql workbench for MySQL and pgAdmin III for PostgreSQL To connect to the databases you use their internal administrative accounts root for MySQL and postgres for PostgreSQL but you must first define its internal postgres password 5 2 Install
30. RTM map names e g the SRTM map N45E0 3 hgt is converted to GRASS map tmp_N45E013 Next the m SRTMtoGRASS sticks merges just converted GRASS maps to the final DEM map using the GRASS r patch command This script requires a Longitude Latitude GRASS location to work and does not run in cartographic projected GRASS locations Parameters n s e w and mapsdir are equivalent to the equally named parameters of m getSRTMmaps e Usage m SRTMtoGRASS p n value s valu value w valu mapsdir string out _DEM string overwrite verbose quiet e Flags p The list of required maps will be printed only no downloading extracting 0 Allow output files to overwrite existing files v Verbose module output q Quiet module output e Parameters n Value for the northern edge degrees 33 s Value for the southern edge degrees e Value for th astern edge degrees w Value for the western edge degrees mapsdir Local directory for SRTM maps out DEM Output DEM e Example m SRTMtoGRASS n 47 0 s 45 3 16 7 w 13 3 out_DEM outdem O The example above merges the SRTM maps covering the Slovenian region downloaded and extracted in the previous step to the final DEM map As a side effect partial DEM maps corresponding to the individual SRTM maps are created 3 4 Project GRASS DEM to the final cartographic location With previous steps we created a DEM map in a Longitude Latitu
31. VM hypervisors e g with VirtualBox The distribution contains three file e ovf file It contains description of the virtual machine number of CPUs memory size and many other things in XML text format e vmdk Virtual Machine Disk file It contains the VM disk image in compressed form This file format was originally developed by VMware but is now an open format e mf Manifest File This file is optional It contains the SHA1 digest of the two previously described files and if present serves to check their integrity The VM disk size is 40 GB the memory size is 1 GB and only on processor one processor core is used The configured memory size and the number of CPU s or CPU cores are very modest so that the machine could be run on a host computer with limited resources This settings can be easily changed and it is recommended to set a larger memory size and more CPU s CPU cores especially if one wants to use the r radcov s parallel execution support The RaPlaT VM was created with software installation and configuration performed according to the installation procedures described in details in the next chapter The following main software was installed e Ubuntu 12 04 LTS Precise Pangolin desktop 32 bit e MySQL 5 5 e PostgreSQL 9 1 e GRASS 6 4 3RC2 a development installation from the source distribution with support for MySQL and PostgreSQL A number of other packages libraries either required or
32. _height value radius value frequency value overwrite verbose quiet e Figas q Quiet 0 Allow output files to overwrite existing files y Verbose module output seg Quiet module output e Parameters inputDEM Name of input raster map clutter Clutter raster map with path loss coefficients output Name for output raster map AO Parameter AO Al Parameter Al A2 Parameter A2 17 A3 Parameter A3 coordinate Base station coordinates ant_height Height of the antenna m default 10 radius Radius of calculation km default 10 frequency Frequency MHz e Example r hataDEM inputDEM dem_slovenia_25 PERMANENT clutter clut_slovenia_25 loss PERMANENT output hataDEM ijs_25 coordinate 460697 99918 ant_height 20 freguency 2000 radius 10 A0 42 A1 42 A2 12 A3 0 1 o 2 2 5 r waik The r waik module implements the Walfisch Ikegami semi deterministic model for path loss computation in microcells It was developed in the framework of the COST231 project 15 and is based on the Walfisch Bertoni 16 and Ikegami 17 models The model computes path loss in two different ways based on LOS Line of Sight It is valid for the receiver and transmitter antenna heights constraints are different for LOS and NLOS cases and are described later e carrier frequency 800 2000 MHz e distance between transmitter and receiver 20 m 5 km e receiver and transmitter height
33. ace the squared distance law 20 dB decade The model ignores terrain configuration relief LOS which is its main drawback and the loss due to land use clutter map The model can give useful results if there are no major obstacles between the receiver and the transmitter An example of a path loss map obtained with r hata is shown in Fig 4 Fig 4 Path loss at 900 MHz computed with r hata e Usage r hata q inputDEM name output name coordinate x y ant _height value radius value area_type string frequency value overwrite verbose quiet 13 e Flags q Quiet 0 Allow output files to overwrite existing files y Verbose module output q Quiet module output e Parameters inputDEM Name of input raster map output Name for output raster map coordinate Base station coordinates ant_height Height of the antenna m default 10 radius Radius of calculation km default 10 area_type Type of area options urban suburban open default urban frequency Frequency MHz e Example r hata inputDEM dem_slovenia_25 PERMANENT output hata_ijs 25 coordinate 460697 99918 ant_height 20 frequency 900 radius 10 o 2 2 3 r cost231 The r cost231 module implements the COST231 empirical model which extends the Okumura Hata model to the 1500 2000 MHz band 15 It is valid for e carrier frequency 1500 2000 MHz e distance between transmitter and receiver 1
34. alled modules RaPLaT reduces the execution times of the path loss models and r sector modules by additionally limiting computation to a circle with a given radius around each antenna The points outside this area are assigned the null value no signal received The r radcov script allows setting the radius independently for each antenna in the input cell list table see below or globally with the radius_ovr command line parameter Radio coverage computation requires a GRASS location cartographic projection with distances expressed in meters e g the Gauss Krueger coordinate system for Slovenia in our case It cannot work correctly in a location with so called Longitude Latitude pseudo projection where locations and distances are expressed in angular degrees 2 1 Run a complete radio coverage computation r radcov A radio coverage computation could be accomplished by calling individual modules described in details later in this document isotropic path loss model modules r sector and r MaxPower Such use would be quite awkward and demanding hence we created r radcov a Python script which ties everything together and calls individual modules as necessary The script gets the necessary information for radio coverage calculations from two tables written in the CSV text format and from the r radcov command line parameters RaPlaT can be used to calculate coverage by radio signals from multiple transmitters as is the case with cellular netw
35. and quick enabling disabling of individual cells An example of radio coverage map received signal strengths in dBm for a system with three antennas on a single location is shown in Fig 2 The antennas used here were not a real product but a mathematically created cosine type with half power 3 dB beam width of about 30 and gain 0 dBd Fig 2 Coverage by three antennas on one location r hata 900 MHz e Usage r radcov rpcl csv_file string antmap_ file string dem_map string clutter map string region string frequency value cellnum value out_map string db driver string database string out_table string dbperf value procnum value model _ovr value radius_ovr value rx _threshold value overwrite verbose quiet e Flags Reuse results from existing intermediate model sector files p purge Delete all unused sector radio coverage files zs check Test run without actually performing radio coverage Computation Sch Rx dBm values in output map replaced by 1 0 when above rx_threshold O Allow output files to overwrite existing files Res Verbose module output q Quiet module output e Parameters csv_file Radio cell sector table in CSV format default sector table csv antmap file Antennas map file default GISBASE etc radio coverage antenna_ diagrams antennamap csv dem_map DEM file for radio coverage simulation default dem_map PERMANENT clutter map Clutter map file required
36. as It does this by taking the path loss raster map s produced by r sector and applying the corresponding transmission power s It obtains the list of all input path loss raster maps with corresponding transmission powers for all the transmission antennas here also called cells in a CSV format text file The module produces a raster map file containing the received strength of the strongest received signal for each raster point If a rx_treshold parameter value is specified the signals with lower received strengths are ignored If flag 1 number one is additionally specified only a simple coverage area map is generated value 1 for the received signal above the threshold 0 otherwise instead of the received signal strength map In addition to the raster map r MaxPower can generate a data table using standard databases supported by GRASS like MySQL or PostgreSQL and also the GRASS own built in DBF containing data about a certain number user selectable parameter cell_num of the strongest received signals in each raster point suitable for further processing by other non GRASS tools The data table generation is activated by specifying the driver parameter with a value other than none dbf for GRASS DBF mysql for MySQL pg for PostgreSQL e Usage r MaxPower ql cell _input string rx_threshold value output name table string driver string database string cell num value dbperf value overwrite verbose quiet
37. ase SQL keywords are usually written in capitals although they are not case sensitive mysql CONNECT grass To list the available databases must be logged in to MySQL SHOW DATABASES To list the available data tables of the connected database SHOW TABLES To exit MySQL g For administrative tasks e g creating new MySQL users the root MySQL login is required no database if connected at login in this example the user will be asked for the password of the MySQL root account 42 mysql u root p The password for root it is normally set during MySQL installation In case of the RaPlaT VM it is the same as for the Ubuntu user account i e user Again it is strongly recommended to change it since any user with an Ubuntu account could potentially login to MySQL as root The MySQL users are stored in the database mysgl in the table user The following commands can be used to list the existing users root login is required the first command lists only the user names while the second one prints all columns of the table SELECT user FROM mysql user SELECT FROM mysql user The MySQL configuration file my cnf is in the etc mysql directory The main database directory is var lib mysql where each database has its own subdirectory for its tables the database grass is stored in the directory var lib mysql grass The MySQL server can be restarted stopped or started with the follow
38. ations with the following commands sudo mkdir var local grassdata cd var local grassdata sudo tar xvzf home userl Downloads RaPlaT_ demolocs tar gz The next step is to set the file and directory ownerships and permissions properly The whole database directory tree ownership is first given to the user user and group grass except for the individual user s mapset where ownerships are given to the corresponding users only the grassuser s mapset in our case sudo chown R userl grass var local grassdata sudo chown R grassuser var local grassdata Slovenia grassuser The above commands differ a bit from those in the Quick Start Guide but the result is the same If we want we can apply some additional settings on the database directory tree chances are that some of these settings are already set by default We can ensure that only the users owners of mapsets have write access on their mapsets sudo chmod R utw go w var local grassdata We can allow non privileged users to freely create their own mapsets in locations by setting the group write permission on location directories sudo chmod gtw var local grassdata It is convenient to set the setgid bit on directories the s command displays it in place of group x permission as s when x is set and S otherwise This way the newly created files and subdirectories will inherit the existing directory group ID grass in our case instead of getting the primary group ID of
39. bove modules The user only uses r radcov and does not need to deal directly with individual modules listed above 5 Some other auxiliary modules C modules and Python scripts Most of the path loss model modules need only DEM Digital Elevation Map Le a raster map describing the terrain profile for their computations Some modules currently only r hataDEM need also a so called clutter map which describes the signal fading at each raster point due to the land use or type of vegetation e g buildings roads forest grass rivers lakes etc Unfortunately geographical maps are generally not publicly and freely available When using RaPlaT in a professional environment e g by a mobile network operator commercial DEM and clutter maps are normally available For non commercial use there are other options GRASS GIS web page itself provides some demo locations in US with DEM and land use maps which can be used for demo purposes Another option are publicly available SRTM Shuttle Radar Topography Mission maps with global Earth coverage which were generated by NASA and are based on radar measurements performed during a Space Shuttle mission in February 2000 6 7 The RaPlaT C modules are currently distributed as source code for Linux environment only They have been tested under Ubuntu 10 04 and 12 04 with GRASS versions 6 4 0 and 6 4 3RC2 Before use the modules must be compiled using standard Linux tools and the GRASS development envir
40. cal tilt value to which the radiation pattern corresponds but is often assigned no value or the keyword ELECTRICAL for antennas not having or having electrical tilt option respectively An actual MSI file could look like this this particular file does not describe a real physical antenna but a mathematically generated one with the cosN radiation pattern NAME COSN21 FREQUENCY 2140 GAIN 19 dBd TILT ELECTRICAL COMMENT simple cos 4 antenna diagram HORIZONTAL 360 0 0 0000 1 0 0139 2 0 0556 357 0 1251 358 0 0556 359 0 0139 VERTICAL 360 O 0 0000 1 0 0139 2 0 0556 35 7 0 1251 358 0 0556 359 0 0139 22 The r sector module only reads and uses the GAIN parameter value expressed in dBd or dBi i e relative to a dipole or isotropic antenna respectively dBd is default x dBd x 2 15 dBi and the pattern definition specified in the HORIZONTAL and VERTICAL sections It calculates the 3 D radiation pattern based on the antenna s given horizontal and vertical patterns its gain and its physical position and direction It then generates an output path loss raster map by applying this pattern to the isotropic path loss raster map previously computed by a path loss model module Fig O shows an example of a path loss map calculated by r sector using the path loss map previously computed by r hata shown in Fig 4 and the just mentioned artificial cosN type antenna The ante
41. cov automatically searches the whole directory subtree for the MSI files The MSI filenames must be unique even if located in different subdirectories 2 1 2 Computation region management In general GRASS modules including the RaPlaT modules perform computations in the so called current region which can be set with the GRASS command g region A region is a rectangle defined by its geographic borders and resolution The r radcov script has a command line parameter called region that allows user to specify a computational region for the radio coverage computation and temporarily sets it as the current region during the computation If signals can be received in this region that emanate from outside transmitters the computational region is automatically enlarged to include those transmitters The region parameter allows setting the computation region in a few different ways e region current the existing current region is used as the computation region this is the default setting e region dem the region of the DEM map its extents and resolution is used as the computation region e region region saved_region_name or region region saved_region_name a prevoiously saved region is used as the computation region the GRASS g region command can be used to save a current region e region rast raster_map or region rast raster_map a GRASS raster map region is used as the computation region region dem can be regarded as a special
42. cy gt H_WIDTH lt h_width gt V_WIDTH lt v_width gt FRONT TO BACK lt front_to_back gt GAIN lt gain gt TILT lt tilt gt POLARIZATION lt polarisation gt COMMENT lt comment gt 21 HORIZONTAL 360 O lt OH gt 359 lt 359H gt VERTICAL 360 O lt 0V gt 359 lt 359V gt The variables are NAME Name of the antenna MAKE Name of the manufacturer FREQUENCY Frequency in MHz H WIDTH Opening angle in the horizontal plane between the 3 dB points V_WIDTH Opening angle in the vertical plane between the 3 dB points FRONT TO BACK Front to back ratio in dB GAIN Antenna gain in dBd when in dBi this must be specified TILT Electrical tilt of the main beam in degrees POLARIZATION Horizontal vertical 45 or 45 COMMENT Comment OH 359H Horizontal gain data points per horizontal angle relative to maximum gain being zero Any value below zero is assumed to be negative Minus sign is not used with these values OV 359V Horizontal gain data points per horizontal angle relative to maximum gain being zero Any value below zero is assumed to be negative Minus sign is not used with these values In practice MSI files usually use only a subset of the parameters listed above e g NAME FREQUENCY GAIN in dBd default or dBi TILT COMMENT and of course the HORIZONAL 360 and VERTICAL 360 sections Besides the TILT parameters does not necessarily specify the actual electri
43. de GRASS location For normal work we usually use GRASS locations with cartographic projections This is especially true for radio coverage computations which require distances expressed in meters Hence the just created DEM map must be projected from the original Longitude Latitude GRASS location to the final projected location This can be done with the GRASS r proj command which maps a raster map from a non current GRASS location to the current one In the following example the LongLat GRASS location is where the original SRTM based GRASS DEM was created Cubic interpolation is used for mapping from the original SRTM original around 90 m in N S direction to the new resolution of 100 m The command is run within the user s final cartographic GRASS location and projects the raster map outdem from the mapset named user in the Longitude Latitude location named LongLat to the currently used final mapset and location r proj input outdem location LongLat mapset userl output outdem resolution 100 method cubic o 3 5 Removing voids As has already been mentioned and will be illustrated in the next chapter for the case of Slovenia SRTM maps and hence the resulting DEM can contain void areas GRASS contains a module r fillnulls that can fill void regions with interpolated values r fillnulls input outdem output outdem_filled This can be done in the final projected GRASS location Of course the interpolated values can differ conside
44. e in degrees non negative values of the lower left S W corner of the 1 degree region E g the file N45E014 hgt which contains a part of Slovenia including its capital Ljubljana covers the region between 45 0 5 and 46 0 5 North latitude and between 14 0 5 and 15 0 5 East longitude For the southern and western location the N and E in the file names are replaced by S and W respectively The filename extension her does not denote any special file format but simply stands for the word height The file contains raw data with no special formatting The data are 1442401 1201x1201 values of elevations above the sea level in m each written as a two byte 16 bit signed integer Voids raster points without a valid height data are indicated by the special value 32768 The integers are written in the big endian form The term big little endian denotes the order of bytes which a particular processor type uses to write multi byte integer values into memory bytes little endian systems store the least significant byte at the lowest memory byte address while big endian systems use the opposite order The Intel x86 processor family uses the little endian form hence in order to read the integer values in hgt files correctly on an Intel based computer architecture independently of the operating system used the two bytes must be swapped Compared to commercial DEM s SRTM maps have some limitations Their resolution is not very
45. ed that builds all the modules in a single step The detailed description is in the RaPlaT Quick Start Guide The RaPlaT distribution is accompanied with two additional archive files one with a demo antenna setup and another with RaPlaT demo commands and CSV files Please consult RaPlaT Quick Start Guide for their description and installation The manual also contains a short description of the demo GRASS database installation which can be used instead of the description in the following subchapter 5 4 Set up GRASS database and users When GRASS is started for the first time it asks for the location of its database with locations and mapsets Hence before starting GRASS we have to set up its database But before doing that we will create an additional Ubuntu group grass all GRASS users should become its members and an additional user account grassuser a generic non privileged GRASS user The following example assumes that the primary user account is user with Ubuntu administrative permissions i e sudo permissions and it will also be used for the GRASS database administration it will be the owner of the database directory tree except for the personal mapsets subtrees which are owned by individual non privileged GRASS users Both user and grassuser will be members of the grass group sudo addgroup grass sudo adduser userl grass sudo adduser grassuser sudo adduser grassuser grass The GRASS DBF database has a form of a direc
46. eeessecsseceseceseceseenseceseeeseescaeecaeeeaaecaaecsaecaeceaeeaeenes 3 2 Coverage by three antennas on one location r hata 900 Ma 6 3 Path loss at 2 GHz computed with fen 12 4 Path loss at 900 MHz computed With 7 botog ce eceesseseeeseeeneecnsecssecnsecnseceseeeseeeseeeneeeeaes 13 5 Path loss at 2 GHz computed With r cost231 eeeecescessecsseceseceseceeceseceeeeeeeeeeaeeeaeeeaeeeaaesnaees 15 6 Basic concept of the hataDEM model 16 7 Path loss at 2 GHz computed with r boroaliEuM AAA 17 8 Path loss at 2 GHz computed with r walk 20 9 Path loss computed by r sector based on r hata path loss Fig Ai 23 10 Official land use map for the Ljubljana regen 28 11 The corresponding clutter map generated by r clutconvert 1cesceseeseceeneeeneeeeeeneeeneeensees 28 12 Commercial Slovenian DEM Google Man 35 13 Official commercial Slovenian DEM 36 14 Slovenian DEM created from SRTM maps ee cecceseceseceseceseceseesseeeseeesaeeeaeeeaeeenaeenaees 36 15 Differences between the comercial and SRTM based Slovenian DEMS sia co sco oo 37 16 Profile difference lower left corner to upper right corner diagonal 4 44 4 4 44 37 17 Profile difference upper left corner to lower right corner diagonal 4 4 4 4 444 37 18 Height difference Nisto grano visitadas deli desidia e DATA DESA PAAU A ADA A De dA Aa 38 19 Slovenian DEM created from SRTM maps with voids filled r fillnulls
47. er in the grass database the new MySQL users userl grassuser can now also use the MySQL client mysql to work with the grass data tables E g to get the list of all data tables in the grass database user would execute the following commands userl ubuntu mysql grass mysql gt SHOW TABLES mysql gt q The default main database directory is var lib mysql where each database has its own subdirectory for its tables the database grass is stored in the directory var lib mysql grass The MySQL configuration file my cnf is in the etc mysql directory If necessary the MySQL server can be restarted stopped or started with the following commands sudo service mysql restart sudo service mysql stop sudo service mysql start 51 5 5 2 Set up PostgreSQL PostgreSQL administration is done using a special Ubuntu user account postgres which is created during the PostgreSQL installation By default no password is defined for this account and direct login is disabled Instead sudo is used and all users having Ubuntu sudo permission have administrative PostgreSQL access A user can become a postgress user with any of the following four commands sudo u postgres s sudo su postgres sudo u postgres i sudo su postgres The last two commands also perform login and set the current directory to the postgres home directory which is the main data directory for PostgreSQL databases Wvar lib postgresql by default Instead of
48. explicitly becoming the postgres user until exit is executed only a per command user change is used in the following examples The grass database can be created with the following command sudo u postgres createdb grass The two users user and grassuser are created with the following commands userl ubuntu sudo u postgres createuser Enter name of role to add userl Shall the new role be a superuser y n n Shall the new role be allowed to create databases y n n Shall the new role be allowed to create more new roles y n n userl ubuntu sudo u postgres createuser Enter name of role to add grassuser Shall the new role be a superuser y n n Shall the new role be allowed to create databases y n n Shall the new role be allowed to create more new roles y n n In the described setup we created no PostgreSQL schema so all users implicitly access the public schema and are granted database access by default Such setup is suitable for one or a few cooperating users in a database Besides being able to create data tables with RaPlaT r radcov r MaxPower in the grass database the new PostgreSQL users userl grassuser can now also use the PostgreSQL client psql to work with the grass data tables To connect to the grass database get the list of all databases data tables users and then exit user or grassuser would execute the following commands command prompts are
49. ficial commercial DEM map 3 2 Download SRTM maps m getSRTMmaps To make downloading the necessary SRTM maps easier we have developed a script m getSRTMmaps It downloads and extracts unzips the SRTM maps required for the chosen region The region is rectangular and defined by its bounds expressed in angular degrees parameters n e s and w In principle m getSRTMmaps does not depend on GRASS and would not need any special GRASS support for it to work However for a consistent user experience the script makes use of the GRASS command line parsing mechanism and hence works only within the GRASS environment i e in a Terminal window where GRASS has been started The GRASS location and its projection properties are not important but since the following steps would normally be creating the longitude latitude DEM map with the m SRTMtoGRASS script described later it would probably be used within a longitude latitude GRASS location The m getSRTMmaps script downloads and extracts unzips maps to a destination directory defined with the parameter mapsdir If mapsdir is not set the current working directory is used During downloading each required zipped map is downloaded only if neither the zipped nor the extracted map with the same name exists yet in the destination directory During extracting each required map is extracted only if no map with the same name exists yet in the destination directory In other words existing file are kep
50. fmpeg yes with ffmpeg includes usr include libavcodec include libavformat usr include libswscale h freetype yes with freetype includes usr include freetype2 h postgres yes h postgresql yes with postgres includes usr include postgresql h sqlite yes h mysql yes with mysql includes usr include mysql h odbc yes Note that the line with ffmpeg above is split into two lines to fit on the page but it should be a single line The warning configure warning libmysqld not found if printed can be ignored If the building process was successful the configuration result is printed in our case GRASS is now configured for 1i686 pc linux gnu Source directory usr local src grass 6 4 3RC2 Build directory usr local src grass 6 4 3RC2 Installation directory S prefix grass 6 4 3RC2 Startup script in directory exec_prefix bin C compiler gcc 03 C compiler c g 02 Building shared libraries yes 64bit support no OpenGL platform X11 MacOSX application no MacOSX architectures MacOSX SDK Tcl Tk NVIZ yes BLAS support no C support yes Cairo support yes DWG support no FFMPEG support yes FFTW support yes FreeType support yes GDAL support yes GEOS support yes GLw support no LAPACK support no Large File support LFS yes Motif support no MySQL support yes NLS support yes ODBC support yes 47 OGR support yes OpenGL support yes PNG support
51. for hataDEM model default clutter _map PERMANENT region Computation region dem current or region rast n e S w res S g xregion default current frequency Radio frequency MHz default 900 cellnum Number of succesive path loss values to be written in the table default 5 out_map Simulated radio coverage raster output default out_raster db driver Database driver options none dbf mysql pg default none database Database name default GISDBASE SLOCATION NAME SMAPSET dbf out_table Simulated radio coverage db table output default out_db dbperf Database insert performance rows INSERT 99 special fast mode via CSV options 1 99 default 20 procnum Number of parallel processes 1 automatic 0 non parallel default 1 model ovr Model override with parameters radius our Radius override km rx threshold Minimum received power dBm for radio signal coverage e Example does not create a data table r radcov csv_file cell list_ijs csv dem_map dem_ slovenia 25 out_map ijs_abc frequency 900 o 2 1 1 Antenna types table The list of available antenna types with corresponding parameters is given in a CSV format file The r radcov script reads the antenna type for each cell from the cell list table CSV file described above and then uses the antenna types table to find the corresponding MSI file describing the antenna s characteristics The MSI
52. format is described later in the r sector chapter The antenna types table file is specified with the r radcov s antmap_file command line parameter The default path is GISBASE etc radio_coverage antenna_diagrams antennamap csv where the GISBASE environment variable is set by GRASS and contains the path to its program directory An example of the antenna types table is shown in Table 3 only one antenna is listed Table 3 An example of the antenna types table antennaType Frequency frequencyLower frequencyUpper EDT MSlfilename technology COS 21 1500 800 2200 0 COS_21 none The CSV file must be in standard format the modified European MS Excel semicolon separated values format is not supported The CSV file corresponding to Table 3 would be generated by OpenOffice Spreadsheet antennaType frequency frequencyLower frequencyUpper EDT MSIfilename technology COS 21 1500 800 2200 0 COS_21 none The first line contains the header each following line contains the following data e antennaType antenna type name allowed character are letters numbers Y and in any order e frequency nominal frequency of the antenna in MHz e frequencyLower the lower frequency bound of the antenna in MHz e frequencyUpper the upper frequency bound of the antenna in MHz e EDT electrical tilt of the antenna in downwards a non negative integer value e MsS
53. ing commands sudo service mysql restart sudo service mysql stop sudo service mysql start Besides the original MySQL user manual 26 short and useful quick start guides can be found in the web e g 27 4 5 About PostgreSQL The PostgreSQL database in the RaPLaT VM already contains a pre created database named grass which can be used to store data tables created by RaPlaT Both preconfigured local users user and grassuser have full access on this database The following commands can be used in a Terminal window to list the existing PostgreSQL databases and data tables the pre created grass database in initially empty To login to PostgreSQL with the database grass psql grass To list the available databases must be logged in to PostgreSQL l To list the available data tables of the connected database dt To list all users SELECT FROM pg _ user To exit PostgreSQL q 43 For administrative tasks the user should run psq as the special Ubuntu user postgres using sudo no postgress password is set and required but only users with sudo permission can run it e g user sudo u postgres psql The PostgreSQL configuration files are in the etc postgresql 9 1 main directory The main database directory is var lib postgresql 9 I main The PosgreSQL server can be restarted stopped or started with the following commands sudo service postgresql restart sudo service postgresql stop sudo
54. ing global DEMs is ASTER 21 however those maps are currently less reliable SRTM maps are an alternative source free for non commercial use to produce a GRASS DEM for a particular geographic region that can be used with RaPlaT for radio coverage computation The suitable version of maps is located at http dds cr usgs gov srtm version2_1 SRTM3 7 It contains six subdirectories for different parts of the words Africa Australia Eurasia Islands North_America and South_America Each SRTM3 map covers a geographic region of 1 degree longitude and latitude with resolution of 3 arc seconds This would result in the raster dimensions 1200x1200 1200 1 3 3600 3 In fact the raster dimension is 1201x1201 with region boundaries extending 1 5 arc second half the pixel size beyond the 1 degree bounds in all for directions E W N S This way the raster pixels on the region borders have their centers exactly at the 1 bounds and theses bordering pixels overlap with the bordering pixels of the four neighboring regions The resolution of 3 corresponds to roughly 90 m on the equator and correspondingly less in the E W longitude dimension at other latitudes Maps are available as zipped files named according to the geographic coordinates latitude longitude in degrees For northern and eastern locations the names have the form NyyExxx hgt with added zip for zipped files where yy and xxx represent the latitude and longitud
55. ion km default 10 free space loss correction Free space loss correction default 32 5 bs correction Reduced base antenna height correction Default 54 range correction Range correction default 10 street width correction Street width correction default 10 frequency correction Frequency correction default 10 building height correction Building height correction default 20 street width Widths of roads m default 15 distance between buildings Building separation default 30 building height Heights of buildings m default 12 PHI Street Street orientation angle deg default 90 area_type Type of area options metropolitan medium cities default medium cities e Example r waik inputDEM dem_slovenia_25 PERMANENT output waik ijs 25 coordinate 460697 99918 ant _height 20 frequency 2000 radius 10 o 2 3 Add transmission antenna r sector The path loss model modules described so far compute path loss for the case of isotropic transmission a hypothetical isotropic antenna with O dB gain without considering actual transmission antenna characteristics position and orientation The next step in to apply the antenna radiation pattern which is the task of the r sector module The antenna radiation pattern and other data must be given in the MSI Planet Antenna File Format 18 This is a text format with the following structure NAME lt name gt MAKE lt make gt FREQUENCY lt frequen
56. le data table in the main memory which quickly runs out for large tables and does not support fast writing modes hence it can be very slow for large tables see also the dbperf parameter GRASS includes support for a number of external databases When building GRASS from source the GRASS support must be activated for each database to be used and that database must already be installed The two recommendable high performance databases that are specifically supported and tested with rMaxPower are MySQL parameter value msyql and PostgreSQL parameter value pg database defines the name of existing database where the output data table will be created The default value represents the location of the GRASS built in DBF database which is always available but is not recommendable for larger radio coverage projects the three environment variables represent GRASS working environment the basic GRASS database folder the user selected GRASS location and the user s GRASS mapset The output data table is stored as a dbf format file named GISDBASE LOCATION_NAME MAPSET dbf lt table_name gt dbf In case 26 of MySQL or PostgreSQL the empty database must first be created with their own tools with an arbitrary name e g a reasonable database name could be grass e table defines the output table name for a particular radio coverage computation e cell_num defines the number of the strongest received signals in each point of the
57. liIfilename the name of the MSI file without the MAT or mei extension that describes the antenna characteristics for this particular combination of antenna type amp frequency amp electrical tilt e technology used to describe the type of the radio communication technology the antennas is made for e g GSM 900 GSM 1800 UMTS 2100 an arbitrary comment not used for processing Empty lines are ignored The character as the first character of the antenna type or the first character in a line has a special meaning it marks the line as a comment only An antenna type can support multiple frequency bands nominal frequencies and electrical tilts with each combination having different characteristics described by the corresponding MSI files Hence the same antenna type can appear in the table multiple times The r radcov searches the table for rows with the required antenna type electrical tilt and with the frequency range defined by the lower and upper frequency bound that includes the simulation radio frequency set by the frequency command line parameter of r radcov If multiple table rows fulfill these requirements r radcov takes the one with the antenna nominal frequency closest to the simulation radio frequency The MSI files must be located in the same directory with the antenna types table file or in any of its optional subdirectories Subdirectories can be used for logically grouping MSI files and have no other meaning The r rad
58. ll_name gt lt antenna_index gt lt sector raster map_name gt lt transmit power gt lt model with parameters gt There can be a single cell or a number of them e g in the case of a cellular radio network No header line is used and no empty lines are allowed including at the end of file The only important columns are lt sector raster map_name gt and lt transmit power gt Other columns are only informal and could be empty they are written to the data table along with the calculated received powers Their purpose is e lt cell_name gt an arbitrary cell name that helps the user identify the cell identity location e lt antenna_index gt antenna index for uniquely identifying each antenna cell in the system there can be more than one antenna in a cell e g two antennas might be connected in parallel to a single transmission signal via a power splitter to obtain a required transmission pattern e lt model with parameters gt contains information about the radio propagation model used and its parameters independently for each antenna Following is an example of a cell list file no empty lines at the beginning end IJS A 1 1IJS A 1 hata urban 460697 99918 20 10 900 30 0 0 COS 21 30 hata urban IJS B 2 1JS B 2 hata urban 460697 99918 20 10 900 135 O O COS 21 30 hata urban TJS C 3 IJS C 3 hata urban 460697 99918 20 10 900 270 O O COS 21 30 hata urban 2 4 2 The output data table The optionally genera
59. mple r fspl inputDEM dem_ slovenia _25 PERMANENT output fspl ijs 25 coordinate 460697 99918 ant _height 20 frequency 2000 radius 10 o 2 2 2 r hata The r hata module implements the Okumura Hata radio propagation empirical model 14 It is one of the most widely used models for radio coverage estimation and is based on the empirically estimated rules measured propagation data It is valid for e carrier frequency 150 1500 MHz e distance between transmitter and receiver 1 20 km e effective BS transmitter antenna height 30 200 m e effective MS receiver antenna height 1 10 m 12 It contains three sub models for urban suburban and open geographic areas as defined by the following equations Ly 69 55 26 16log f MHz 13 82log h m Cy 44 9 6 551og h m log R km 2 Lu Ly tog RE 5 4 3 28 Lo L 4 78 log f MHz 18 33log f MHz 40 94 4 C 0 8 1 1log f MHz 0 7 h m 1 561og f MHz 5 where Ly Lsu Lo loss in dB for urban suburban and open environments respectively h difference between the transmitter and receiver antenna heights hy receiver antenna height above the ground Co correction factor related to the receiver antenna height R distance between the transmitter and the receiver f transmission frequency in MHz The rate of signal loss with the distance depends on the antenna height For a very high antenna it approximates the loss in empty sp
60. nd The result is shown in Fig 19 Voids are gone but while this can work well for small areas large height errors can be introduced for large void areas as can be seen in Fig 20 some mountain peaks are completely lost with the interpolated heights up to 1259 m below the actual peak height Fig 19 Slovenian DEM created from SRTM maps with voids filled r fillnulls 38 Fig 20 Differences between the comercial and the SRTM based voids filled Slovenian DEMs Hence filling large voids with interpolation is problematic and could cause completely false radio coverage results in case of putting a transmitter on a such location 39 4 Virtual machine GRASS amp RaPlaT preinstalled To make the use of GRASS RaPlaT easier without the need to first install everything from scratch we prepared a VM Virtual Machine a virtual appliance with Ubuntu OS and GRASS RaPlaT tools installed It was prepared with WMware Player which is a simple VM hypervisor meant for personal desktop use It is free for non commercial use and can run on Linux and MS Windows operating systems 8 The original VM would run natively in VMware Player and could easily be converted with VMware tools for other VMware s VM hypervisor e g for ESXi which is used in multiple user server based environments However for distribution we converted the RaPlaT VM to the OVF format Open Virtualization Format which is compatible also with other vendors
61. nna is directed 30 eastwards north is the reference positive values correspond to the clockwise rotation Fig 9 Path loss computed by r sector based on r hata path loss Fig 4 e Usage r sector q pathloss raster name inputDEM name output name ant_data_file string beam_direction value mech _tilt value height _agl valu ast value north valu radius value overwrite verbose quiet e Flags q Quiet SO Allow output files to overwrite existing files V Verbose module output g Ouiet module output e Parameters pathloss raster Omni antenna path loss raster input DEM Elevation model required for transmitter height determination output Name for output raster map ant data file Antenna data file beam direction Beam direction deg mech tilt Mechanical antenna tilt deg height_agl Above ground level height m east Easting coordinate north Northing coordinate radius Radius of calculation km default 10 23 e Example r sector pathloss raster hata_ijs_ 25 inputDEM dem_ slovenia 25 output sector hata ijs 25 ant data file usr local src grass 6 4 3RC2 dist i686 pc linux gnu etc radio_coverage antenna_diagrams demo COS 21 MSI beam direction 30 mech _tilt 0 height agl 20 radius 10 east 460697 north 99918 o 2 4 Calculate complete coverage r Max Power The r MaxPower module calculates the received radio signal strength s from one or more transmitters transmit antenn
62. ny case the previously created binaries created within usr local src grass 6 4 3RC2 directory tree would remain unchanged Finally we can update links to the most recent shared libraries with sudo ldconfig Before starting GRASS for the first time its DBF database directory must be set up see the next chapters 5 3 Install RaPlaT modules The RaPLaT modules C modules and Python scripts are distributed as source code each in its own directory These directories should be copied into doc raster for C modules and doc python for Python scripts subdirectories of the usr local src grass 6 4 3RC2 directory We then cd to each of the modules directories and execute make For C modules this compiles the modules and generates the binary code in the bin subdirectory within dist i686 pc linux gnu the proper place for GRASS binaries and html help files in the docs html 48 subdirectory For Python scripts make copies scripts to the appropriate place i e the scripts subdirectory within the usr local src grass 6 4 3RC2 The make process for the Python scripts is not really necessary since they are already executable and we can just copy them manually wherever we want From now on the RaPlaT modules and scripts can be called like any GRASS module or script within the GRASS environment i e in the Terminal windows where GRASS has been started and is running To simplify the installation a shell script RaPlaT_make has been prepar
63. oints where its received strength in dBm falls below the threshold value The flag number one modifies the output coverage map Normally the map contains the received strengths of the strongest received signal in each raster point in dBm When this flag is specified usually together with rx_treshold the map contains only binary information about the signal coverage 1 e 1 0 or 0 0 if a point is covered or not covered respectively by a signal above the threshold if one is specified The flag c check causes r radcov to not call and execute modules Instead it only prints all the commands module calls that would be executed and the contents of the related input cell list file generated by r radcov for r MaxPower The file itself is a temporary file and is automatically deleted when r radcov completes its execution 2 2 Radio propagation models isotropic antenna RaPLaT containes a number of modules that calculate radio signal path loss according to various path loss models The result is a raster GRASS map with each point having the value of the signal fading in dBm in that point relative to the transmitter no particular antenna is assumed yet the situation corresponds to the isotropic radiation diagram with 0 dB gain 2 2 1 r fspl The r fsp module calculates the radio signal loss according to the free space model FSPL Free Space Path Loss according to the equation 1 13 FSPL 32 4 20log R km
64. om doc refman 5 5 en index html http downloads mysql com docs refman 5 5 en a4 pdf MySQL MySQL Quick Start Guide http www coderguide com Guides MySQL MySQL _ Quick Start Guide PostgreSQL 9 1 9 Documentation The PostgreSQL Global Development Group http www postgresql org docs manuals http www postgresql org docs 9 1 interactive index html http www postgresql org docs 9 1 static index html http www postgresql org files documentation pdf 9 1 postgresql 9 1 A4 pdf PostgreSQL QuickStart Reference Commands http www linuxweblog com postgresql reference GRASS Wiki Compile and Install Ubuntu http grasswiki osgeo org wiki Compile_and_Install_Ubuntu GRASS GIS Download http grass osgeo org download http grass osgeo org download software g64x 55
65. onment Standard precompiled GRASS Linux packages do not include the development support so GRASS must be installed from its source code distribution For someone familiar with Linux this is not a big problem but to make RaPlaT more accessible for users with less Linux experience we created a virtual machine with Ubuntu 12 04 and all the required components preinstalled The virtual machine was prepared with WMware Player which is free for non commercial use and can run on Linux and MS Windows operating systems 8 2 RaPlaT in details The main overall structure of the RaPlaT tools is depicted in Fig 1 It consists of a number of path loss model modules implementing different radio propagation models the r sector and r MaxPower modules and the r radcov Python script which ties everything together Input and output data are depicted in Fig 1 as differently colored parallelograms textual input and output files in orange GRASS raster files in blue and databases in yellow The user can call individual modules however he she would normally only call r radcov which in turn calls other modules as necessary The user defines the parameters of one or more radio transmitters together with the chosen path los models in a cell list file which is a simple data table in the CSV Comma Separated Values format 9 10 The list of all available antenna types is given in another CSV format file which references the actual antenna data files written in
66. onment and Python scripts They belong to the following groups 1 A group of path loss model modules each calculating radio signal path loss according to a specific radio signal propagation model The obtained raster diagram which tells the path loss in dB in each point of the terrain surface corresponds to a hypothetic isotropic transmission antenna with 0 dB gain This group currently comprises the following modules e r fspl Free Space Path Loss model e r hata Okumura Hata model e r cost231 COST 231 model e r hataDEM modified Hata Okumura Hata DEM model e r waik Walfish Ikegami propagation model 2 Module r sector which takes the isotropic path loss results calculated by a path loss model module and modifies it according to the selected antenna characteristics radiation pattern and gain its position and orientation 3 Module r MaxPower which calculates the received power in each raster point of the terrain surface for one or more transmission antennas e g for a cellular communication network like GSM or UMTS In case of multiple transmission antennas it calculates the maximum received power from any transmitter in each receive point raster point on the terrain relief map and can also build a data table using e g MySQL or PostgreSQL comprising the relevant data of a chosen number of strongest received signals in each receive point 4 Script r radcov which helps the user by automatically calling the a
67. ons MSI Digital elevation data http www viewfinderpanoramas org dem3 html CGIAR CSI SRTM 90m Digital Elevation Database v4 1 http www cgiar csi org data srtm 90m digital elevation database v4 1 54 21 22 23 24 25 26 27 28 29 30 31 ASTER Global Digital Elevation Map Announcement NASA Jet Propulsion Laboratory http asterweb jpl nasa gov gdem asp Shuttle Radar Topography Mission SRTM mission statistics NASA Jet Propulsion Laboratory http www 2 jpl nasa gov srtm statistics html Rodriguez E C S Morris J E Belz E C Chapin J M Martin W Daffer S Hensley 2005 An assessment of the SRTM topographic products Technical Report JPL D 31639 Jet Propulsion Laboratory Pasadena California 143 pp http www2 jpl nasa gov srtm SRTM_D31639 pdf Ricardo Passini Karsten Jacobsen Accuracy analysis of SRTM height models http www ipi uni hannover de uploads tx_tkpublikationen RP_KJ_07_SRTM pdf Antonios Mouratidis Pierre Briole Kostas Katsambalos SRTM 300 DEM versions 1 2 3 4 validation by means of extensive kinematic GPS measurements a case study from North Greece International Journal of Remote Sensing Vol 31 No 23 10 December 2010 6205 6222 http users auth gr kvek mouratidis_et_al_2010_ijrs pdf MySQL 5 5 Reference Manual Including MySQL Cluster NDB 7 2 Reference Guide http dev mysql com doc index html http dev mysql c
68. orks e g GSM radio network The user describes the whole configuration in a cell list file cell list here is actually a list of installed antennas with related data as will be explained shortly The file in the CSV format and can be created with OpenOffice Spreadsheet but also with MS Excel in the MS Windows environment r radcov understands the peculiarities of the MS Excel CSV format including its European version the RaPlaT tool itself is currently supported only on Linux The cell list file is specified with the csv_file command line parameter and contains a table an example of which is shown in Table 1 three transmit antennas on a single location Table 1 An example of the cell list table E pe o E U O o sIZ 3 S 5 w Es S 5 S Q DE lo E U E i E Ea S S g 5 5 z y E E E 2 2 o a E A Sigill S s E E 2 8 Ee 8 s s of e 2 s 2 2 2 a 8 0 Ge LSA 1 30 0 0 20 COS 21 460697 99918 30 10 hata urban ef lJS B 2 135 0 0 20 COS 21 460697 99918 30 10 hata urban S S LSC 3 270 0 0 20 COS 21 460697 99918 30 10 hata urban S S The corresponding CSV file as generated by OpenOffice Spreadsheet would be cellName antennaID beamDirection electricalTiltAngle mechanicalAntennaTilt heightAG L antennaType positionEast positionNorth power radius model P1 P2 P3 p
69. ormation about the signal loss at the receiver location related to the land use e g urban agricultural forest areas etc This information is given in the form of a raster map called clutter map where the value of each raster point specifies the received signal loss in dB in that point due the land use Land use is often given as a raster map with values specifying the use e g 1 irrigated agriculture 2 rangeland 3 coniferous forest 4 deciduous forest 5 mixed forest 6 disturbed The r clutconvert module converts such land use map to the corresponding clutter map The conversion from land use types integer numbers to signal loss values in dB is specified by a table defined in a text file The numbers representing land use depend on the particular land use map and the radio signal loss values depend on the radio frequency Hence even if there is only one land use map there will probably be a number of these conversion table files each for a particular frequency band The one to be used by r clutconvert is specified with its Path_loss_values command line parameter The conversion table file consists of a number of lines Lines can be either comment or data lines A comment line starts with and is ignored A data line specifies conversion from a land use value an integer number to the corresponding radio signal loss in dB generally a floating point value with as separator The following is an example of this
70. ose module output q Quiet module output e Parameters inputDEM Name of input raster map output Name for output raster map coordinate Base station coordinates ant_height Height of the antenna m default 10 radius Radius of calculation km default 10 area_type Type of area options medium_cities metropolitan default medium_cities frequency Frequency MHz e Example r cost231 inputDEM dem_ slovenia _25 PERMANENT output cost231_ ijs 25 coordinate 460697 99918 ant _height 20 frequency 2000 radius 10 o 2 2 4 r hataDEM The r hataDEM module implements a modified extended Okumura Hata model The radio signal loss depends on the transmission radio frequency the distance between the transmitter and the receiver the height of the transmit and receive antennas and also on the terrain profile land use and earth surface curvature The model is valid for e carrier frequency 10 MHz 2 GHz 15 e distance between transmitter and receiver 200 m 100 km e effective BS transmitter antenna height 20 200 m e effective MS receiver antenna height 1 5 m The basic concept of the model is shown in Fig 6 Parameters Hm Hb F AO A3 Terrain profile A E model hataDEM Path loss surface curvature Constants Land use related loss Fig 6 Basic concept of the hataDEM model The general path loss equation of the model is L dB HOA dB mk J aKDFR JDFR 8 where HOA Okumura Hata
71. output coverage raster that are stored in the data table see Table 5 e dbperf stands for database performance and selects faster modes of writing tables For dbperf values between 2 and 98 the so called multiple row inserts are used where instead of a single data row a group of data rows 2 to 98 is inserted with each SQL INSERT statement For MySQL and PostgreSQL a reasonable value is 20 which is the default In our case it speeds up the data table creation around 2 5x for PostgreSQL and 3 8x for MySQL relative to the basic single row insert mode For the GRASS built in DBF database dbperf value should be set to 1 the basic non accelerated mode because GRASS DBF does not support multiple row inserts The value of 99 selects a special very fast mode which is supported only for MySQL and PostgreSQL It doesn t write data rows directly into the output data table but instead creates an intermediate CSV text file This is at the end converted to the data table in a single step using non standard database specific commands for MySQL LOAD DATA LOCAL INFILE file csv INTO TABLE lt table gt FIELDS TERMINATED BY LI ENCLOSED BY for PostgreSQL COPY lt table gt FROM file csv CSV QUOTE In our case by using this mode speeds up of data table creation of around 20x were achieved 2 5 Prepare clutter map r clutconvert In addition to DEM some radio signal propagation models hataDEM in our case need also inf
72. port sudo apt get install libmysqlclient dev 2 Install pre compiled packages for PROJ 4 GEOS and GDAL of course this can also be done with the graphical Ubuntu Software Center instead of the following commands sudo apt get install libproj dev sudo apt get install libgeos dev sudo apt get install libgdal dev 3 Install GRASS From 31 download the GRASS source in our case the latest version 6 4 3RC2 Extract the downloaded file grass 6 4 3RC2 tar gz to usr local src take ownership of the resulting directory grass 6 4 3RC and make it the current director cd cd usr local sre sudo tar xvzf home userl Downloads grass 6 4 3RC2 tar gz sudo chown R userl userl grass 6 4 3RC2 cd usr local src grass 6 4 3RC2 46 Run configure with suitable configure parameters for the GRASS build process type configure help to print help on usage and options in the following example support for PostgreSQL is enabled and some non default locations are specified CFLAGS 03 LDFLAGS s configure enable largefile yes wi wi wi wi wi wi wi wi wi wi wi wi usr wi wi wi wi wi wi e CT EE eh ote ect CEET EL EE CC Ch ct ct h readline h nls h cxx h proj share usr share proj h gdal h geos usr bin geos config h python yes h wxwidgets h cairo h tcltk includes usr include tcl8 5 h opengl libs usr include GL h f
73. ps by using the mouse scroll wheel 4 1 Preconfigured Ubuntu user accounts The following Ubuntu user accounts are preconfigured e user this is the primary administrative account with sudo rights It is also the owner of the main GRASS database directories and the PERMANENT subdirectories hence this user is also the GRASS administrator e grassuser this is an ordinary system account but it is also the owner of the corresponding GRASS mapset subdirectory s This account can serve as a model for an ordinary GRASS user The initial passwords for both accounts are equal to the usernames It is strongly recommended to change both as soon as possible especially when and before enabling the remote access over the network Both user accounts are members of the Ubuntu grass group specially created for GRASS users A user must be a member of the grass group to have access to the GRASS database containing GRASS locations with users s mapsets 4 2 About GRASS installation The whole GRASS installation procedure would create two sets of binaries The first one is generated during the compilation make within the GRASS source directory created by decompressing the GRASS source distribution In our case the GRASS source directory is usr local src grass 6 4 3RC2 the binaries are located in the dist i686 pc linux gnu subdirectory and the command shell script grass64 that is used to start the GRASS application is located in bin i68
74. rably from the correct ones especially if the void regions are large 3 6 What about land use clutter maps Certain models e g hataDEM require a clutter map land use related loss maps in addition to DEMs In order to use such models we must provide a clutter map It can be generated from a land use map with the previously described r clutconvert GRASS RaPlaT module In absence of a real land use map a fake clutter map can be prepared that defines a suitable constant land use related loss over the whole region covered by DEM It can be generated by the GRASS r mapcalc command Since it works in the current GRASS region we must first set the region to that of DEM to produce a region wise compatible clutter map 34 The following commands would produce a clutter map named outclut with 10 dB loss over the whole DEM region defined by the raster map named outdem g region rast outdem r mapcalc outclut 10 0 3 7 An example SRTM based DEM for Slovenia With the tools and procedures described in the next chapters we will generate DEM for Slovenia and compared it to the official commercial DEM We will use the following GRASS region definition for Slovenia as printed by g region GRASS 6 4 0 Slovenia gt g region p projection 99 Transverse Mercator zone 0 datum towgs84 426 9206 142 6186 460 0871 4 90806 4 488093 12 423166 17 1128 ellipsoid bessel north 195000 south 25000 west 370000 east 630000 nsres 100
75. s e g the Mercator projection with its parameters suitably defined for a particular location Positions on these maps can be expressed as distances in meters relative to a suitably chosen reference point Different countries generally use different cartographic projections In GRASS a basic property of each GRASS location which in turn includes users GRASS mapsets is its cartographic projection which is defined at the time of the location creation The positions on maps belonging to such projected locations are expressed in meters relative to the corresponding reference point An example of such a GRASS location projection that we use for Slovenia is as printed by g proj GRASS 6 4 0 Slovenia gt g proj p PROJ_INFO name Transverse Mercator proj tmerc ellps bessel lat_0 e P lon O 1715 k 20 9999 x D 500000 y O 5000000 towgs84 426 9206 142 6186 460 0871 4 90806 4 488093 12 423166 17 1128 no_defs defined SE RENE EE EE EE unit metre units metres meters SA GRASS 6 4 0 Slovenia gt A GRASS location can also be created without a cartographic projection by choosing the so called Latitude Longitude pseudo projection The positions on the corresponding map are then expressed in angular degrees The projection properties of such location are as printed by g proj GRASS 6 4 0 LongLat gt g proj p PROJ_INFO
76. s password This password is usually defined during the MySQL installation and is independent from the Ubuntu root password which is by default not defined and login is disabled The following command sequence can be used to create MySQL users and grant them the necessary permissions on the grass database The first command runs the MySQL client performing login to the MySQL administrative root account The second command lists existing databases and is used to verify that the grass database does exist The following four commands create two MySQL local users user and grassuser and grant them all permissions on the data tables in the grass database The last command ends the MySQL client session Command prompts are shown to differentiate between the Ubuntu shell and PostgreSQL client commands userl ubuntu mysql u root p mysql gt SHOW DATABASES mysql gt CREATE USER userl localhost mysql gt CREATE USER grassuser localhost mysql gt GRANT ALL ON grass TO userl localhost mysql gt GRANT ALL ON grass TO grassuser localhost mysql gt q The MySQL users are stored in the table user of the database mysql The following commands can be used to list the existing users the first command lists only the usernames SELECT user FROM mysql user SELECT FROM mysql user Besides being able to create data tables with RaPlaT r radcov r MaxPow
77. t i unconstrained Antenna position E W in m positionNorth i unconstrained Antenna position N S in m power f 0 0 140 0 Transmission power in dBm mW 10kW radius f 0 0 1000 0 Max distance of the receiver in km model s hata cost231 Radio signal path loss model hataDEM waik fspl itm Parameters P1 P11 for the Hata model P1 s urban suburban Area type for the Hata model open P2 P11 not used Parameters P1 P11 for the Cost231 model P1 s metropolitan Area type for the Cost231 model medium cities P2 P11 not used Parameters P1 P11 for the hataDEM model P1 f Unconstrained Parameter AO for the hataDEM model P2 f Unconstrained Parameter A1 for the hataDEM model P3 f Unconstrained Parameter A2 for the hataDEM model P4 f Unconstrained Parameter A3 for the hataDEM model P5 P11 not used Parameters P1 P11 for the Walfisch Ikegami waik model P1 f 20 60 Parameter WO Free space loss correction P2 i 30 70 Parameter W1 Reduced base antenna height correction P3 i 5 35 Parameter W2 Range correction P4 i 3 15 Parameter W3 Street width correction P5 i 3 25 Parameter W4 Freguency correction P6 i 10 30 Parameter W5 Building height correction P7 i 10 25 Parameter W6 Street width m P8 i 20 50 Parameter W7 Distance between buildings m P9 i 0 300 Parameter W8 Building height m P10 i 0 180 Parameter P
78. t unchanged and the corresponding download extract operations are skipped with informative message printed on the screen By default m getSRTMmaps seeks the required maps in the Eurasia subdirectory http dds cr usgs gov srtm version2_1 SRTM3 Eurasia For other geographic regions the 32 user must explicitly specify URL with the urldir parameter substituting Eurasia with Africa Australia Islands North America or South America e Usage m getSRTMmaps p n value s valu value w value urldir string mapsdir string verbose quiet e Flags p The list of required maps will be printed only no downloading extracting CM Verbose module output q Quiet module output e Parameters n Value for the northern edge degrees s Value for the southern edge degrees e Value for th astern edge degrees w Value for the western edge degrees urldir URL of the required SRTM maps default http dds cr usgs gov srtm version2_ 1 SRTM3 Eurasia mapsdir Local directory for SRTM maps e Example m getSRTMmaps n 47 0 s 45 3 16 7 w 13 3 The example above downloads and extracts SRTM maps for the Slovenian geographic region 3 3 Convert SRTM maps to a GRASS DEM m SRTMtoGRASS The m SRTMtoGRASS script first converts the required previously downloaded and extracted SRTM maps to GRASS maps using the GRASS r in bin command The GRASS map names are generated by adding prefix tmp_ to the original S
79. tations it is possible to execute multiple modules in parallel on a multi core processor 11 12 To speed up coverage computation for multiple antenna communication networks r radcov is capable of calling and executing modules in parallel The number of concurrently executing modules path loss model modules in the first step and r sector in the second step is defined by the value of the procnum command line parameter By default or setting procnum 1 automatic mode the number of concurrently executing modules equals the number of existing processor cores in the system A positive value explicitly defines the number of modules to be executed in parallel Even if it is set to 1 the underlying parallel scheduling and stdout stderr buffering mechanisms are still active The parallel mode of execution and related supporting mechanisms can be switched off completely by setting procum 0 2 1 4 Reuse and Purge During the execution of r radcov intermediate GRASS maps are created by the path loss model modules and by r sector They are not deleted automatically and can optionally be reused for another similar coverage computation these can be considered as a kind of caching of the intermediate results for future computations By avoiding unnecessary and possibly lengthy re computations of model path loss and r sector intermediate results the overall time required for a radio coverage computation can be reduced considerably The names of intermedia
80. te maps are generated automatically and contain important information that r radcov needs to reuse them automatically in another radio coverage computation The names of the maps generated by the path loss model modules are built according to the following pattern model Pl_ Pmax _positionEast_ positionNorth_heightAGL radius frequency e g hata urban 460697 99918 20 10 900 The names of the maps generated by r sector are based on the above pattern model in the pattern below extended with additional r sector related information cellName antennalID model beamDirection electricalTiltAngle mechanicalAntennaTilt_antennaType e g IJS A 1 hata urban 460697 99918 20 10 900 30 O O COS 21 By default r radcov ignores any existing intermediate maps and calculates everything anew The user can request reusing existing maps by specifying flag r reuse however this must be done with great caution The intermediate map names do not contain information about the DEM and clutter maps and the computation region that were used during their creation Hence the user must keep in mind that the cached maps may only be used if the maps and the computation region did not change The number of intermediate maps can become quite large making a mess out of the user s mapset When not needed any more the user can request r radcov to delete them by specifying flag p purge all maps with the names following the above patterns will be
81. ted data table contains one row of data for each raster point of the output coverage raster map Raster points with no coverage no received radio signal from any transmitter are not included in the map nevertheless the table can be quite large and needs considerable time for creation The row format of the table is shown in Table 5 Table 5 Output data table format Column x y Resolution cell1 id1 Pri modeli Gell idN PrN model EN name TOUT int int int varchar int real varchar varchar int real varchar real 6 6 4 32 6 6 128 32 6 6 128 6 Description of columns 25 X x coordinate of geographic location a map raster point in m integer 4 bytes default print length is 6 y y coordinate of geographic location a map raster point in m integer 4 bytes default print length is 6 resolution resolution of the computed coverage map in m integer 4 bytes default print length is 4 celli cell name varchar max 32 chars bytes 5 idi antenna index integer 4 bytes default print length is 6 6 Pri calculated received signal power in dBm received in this geographic location output map raster point from cell real 4 bytes floating point default print length is 6 Valid values are gt 999 0 The value of 999 0 has a special meaning no signal received replacing oo modeli Path loss model name
82. tory tree with the first level subdirectories representing locations and the second level directories representing user s mapsets within a location A special mapset PERMANENT is present in each location and is readable by all GRASS users and contains the location s shared maps Before starting GRASS only the main database directory must exist A suitable location for it would be var local e g var local grassdata After starting GRASS the administrative GRASS user the one with write access in this directory can than create a new location using the Location wizard button the special PERMANENT mapset is created in the location and all users can create their mapsets in a location using the Create mapset button provided that the write permission on the location directory is set for the GRASS users group grass in our demo case Instead of starting with an empty GRASS database it might be helpful to set up the demo GRASS database distributed with RaPlat This database includes two locations e Slovenia which can be used for radio coverage computations It includes mapsets for two users user the GRASS administrator and grassuser a normal non privileged GRASS user e LongLat which can be used to convert and join SRTM maps to a GRASS DEM maps It has only the PERMANENT mapset for use with user1 49 The GRASS administrator user in our case can create the database directory var local grassdata and extract the RaPlaT demo loc
83. with parameters used for this celli varchar max 128 chars bytes E No the received power od the strongest signal divided by the sum of the received powers of all signals in dB real 4 bytes floating point default print length is 6 Valid values are gt 999 0 The value of 999 0 has a special meaning no signal received This data is useful for single frequency band systems such as those using CDMA e g UMTS Columns 4 5 6 and 7 repeat N times as defined by the command line parameter cell num but not exceeding the number of all cells specified in the input cell list file The command line parameters related to the output data table are driver defines the database management system used called also simply the database The default value none means that no database is used and no data table is generated other parameters are ignored but dummy values for the table and cell_num parameters must be specified anyway Contrary to r radcov which does not check for actually installed and supported databases but have a fixed list of them r MaxPower uses GRASS s information about available databases and lists them when called with flag help Value dbf selects GRASS own built in database which GRASS uses for its own purposes It is a relatively simple database with limited functionality and efficiency and is not really recommended for use with RaPlaT except for simple cases Before writing to the disk it creates the who
84. yes PostgreSQL support yes Python support yes Readline support yes SQLite support yes Tcl Tk support yes wxWidgets support yes TIFF support yes X11 support yes Now we can proceed with compiling GRASS using the configuration from the previous step make j2 If the process is successful GRASS binary distribution subdirectory dist i686 pc linux gnu or similar is created in usr local src grass 6 4 3RC2 This distribution is fully functional without any further installation steps This location is also the location where third party modules in our case the RaPlaT modules can be compiled as described later GRASS can be started by running the shell script grass64 in the bin i686 pc linux gnu subdirectory To make this script callable without explicitly specifying its location we create a symbolic link grass64 d in usr local bin sudo ln s usr local src grass 6 4 3RC2 bin i686 pc linux gnu grass64 usr local bin grass64 d Note that the line above is split into two lines to fit on the page but it should be a single line From now on GRASS can be started simply by typing grass64 d in a Terminal window of course some other name can be used instead There is an optional additional step in installing GRASS which we will not perform We could install GRASS to the standard Ubuntu system location for binaries with sudo make install or we could create a standard binary installation package with sudo checkinstall In a

Download Pdf Manuals

Related Search

Related Contents

Avaya 2400 DCP Phone Series (2402,2410 and 2420) User's Manual    使用方法 使用说明 사용 지침 Cardiac Output Cables  STR-DE685  User Manual PZ249E - Physik Instrumente  Acer AcceleRAID 352 Network Card User Manual  ノヴェア - CLUB T-fal  manuali  INS & USER GUIDE TALLATION MANUAL  Zenoah HEZ2610F User's Manual  

Copyright © All rights reserved. Failed to retrieve file