UMTS Network Radio Dimensioning & Planning Dr. Hatem MOKHTARI UMTS Senior RF Engineer
Presentation plan • Introduction • Radio design consideration and assumptions — Differences between GSM and UMTS design, UMTS concepts — Services, traffic forecast , areas to be covered, quality of coverage
• Radio dimensioning — Link budget, cell count
• Radio planning — Radio planning tool — Coverage analyses, Monte Carlo simulations — Maps & Statistics — RF design optimisation
• Conclusions
1
Introduction • Designing a UMTS radio network needs to take into account the UMTS features: — UMTS is a Wideband CDMA system – Coverage and capacity are closely correlated
multi-service system — UMTS is a multi-service – Voice and different data services share the same radio resource
• Before starting the UMTS network design, the design objectives should be clearly stated: — Services to be offered — Multi-service network capacity
Multi-service coverage areas — Multi-service — Quality of coverage of each service 2
UMTS Design considerations GSM Service Types • GSM — Voice — SMS — Circuit switched data (9.6kbps/14.4kbps) — Packet switched data (GPRS/E-GPRS)
• GSM is principally designed for voice traffic today.
• There is a need to revisit GSM network designs to quantify performance of data services i.e. GPRS and EDGE
3
UMTS Design considerations (cont.) UMTS Service Types • UMTS — Voice — SMS — Circuit switched data (nominally to 2 Mbps) — Packet switched data (nominally to 2 Mbps)
• UMTS networks will be designed primarily for a mix of data services
• It is forecasted that there will be 12 million subscribers to 384kbps data services by 2005 (UMTS Forum Report #5)
4
UMTS Design considerations (cont.) What’s the Difference? • GSM — Shared radio resource: channels can be dimensioned separately – e.g. – e.g. partitioned voice, GPRS data resources — Primarily voice service based design, with planned terminal capabilities to support only moderate rate data services. — Quality of channels impacted by inter-cell interference.
5
UMTS Design considerations (cont.) What’s the Difference? (cont.) • UMTS — Shared Shared radio radio resource: resource:-- channels channels must must be jointly dimensioned – i.e. mutual interference impacts capacity Signal SNR
— Quality and capacity impacted by intracell and inter-cell effects — True wireless data system with terminals expected to offer multiple services up to typically 384kbps. – Packet switched services to dominate
Signal after de-spreading
Noise includes every other channel on carrier
6
UMTS Design considerations (cont.) Radio design • GSM — Linear process – Coverage, Capacity and Parameter planning largely independent and sequential — Static modeling adequate for both initial and detailed planning
• UMTS/WCDMA/CDMA — Nonlinear process – Coverage, Capacity and Parameter Planning highly interrelated — Static modeling for ‘first pass’ initial design for new ne w entrants — Dynamic modeling for detailed design
7
UMTS Design considerations (cont.) UMTS/CDMA concepts • Cell range & cell capacity are limited by — interference in uplink — power in downlink
• Cell breathing
8
UMTS Design considerations (cont.) UMTS/CDMA concepts (cont.) • Interference limited system • Ec/Io determines the coverage performance, NOT signal strength
P
INTERFE INT ERFERENCE RENCE
P
USEFUL SIGNAL
f
f Before spreading
After spreading
9
UMTS Design considerations (cont.) UMTS/CDMA concepts (cont.) P
P
f
f After de-spreading
After band pass filter
SIR_after_de-spreading (Eb/No) = SIR_before_de-spreading * Processing_Gain
- Eb/No is the main factor that determines call performance such as FER
- Hand Handof offf par param amet eter ers s are are bas based ed on on Ec/ Ec/Io Io val value ue
10
UMTS design process • Iterative
Design Targets
— considering multiple service usage across region
• Nonlinear
Coverage Planning
Capacity Planning
Parameter Planning
— Coverage planning – Geographical service availability — Capacity planning – Ensuring sufficient capacity in network — Parameter planning – Network optimization
11
PlaNet Tool and Database • PlaNet tool:
Planet version 2.8 with W-CDMA module from MSI • PlaNet database 20m resolution clutter and terrain database
Hardware platforms (used by Nortel): UNIX environment: Sun UltraSparc1 with 256 MB RAM, 4 GB hard drive, dri ve, and Solaris 2.5.1. Operating System
PC environment: Dell Precision 410 with 366+ MHz processor, 1 GB RAM, 9 GB hard drive, and Windows NT 4.0 with Service Pack 4
• Over 10 years of experience (GSM, TDMA, CDMA…) 12
UMTS Design • Suggested UMTS Application with Planet 2.8: • UMTS Coverage design is based on PCCPCH, it is handled by Planet as a regular Pilot Channel • Overall power required by UMTS overhead channels is added together under a single Planet defined overhead channel • Forward link power range for traffic channel is set for highest desired data rate at the cell edge • Cell count design has to be based on this data rate i.e. most constraining service
13
UMTS Link Budget (Generic)
• One link budget - per morphology - envi enviro ronm nmen entt - cove covera rage ge typ typee - per service • Only the most constra-ining service in each morphology was used for link budget calculation • Important parameters: Base station noise figure Reverse Eb/No Building penetration loss Area and cell edge reliability Fade margin Handoff gain
LINK B UDGET General
Spreading b andwidth (kHz) (kHz) Thermal noise no ise (kTB) (kTB) (dBm) (dBm) Data rate (kbps ) Bearer rate rate (kbps ) Process ing gain ga in UPLIN UPLINK K Processing gain (dB) UPLINK UPLI NK Characteristics
Macro cell Indoor Suburban S3 111 3840 -108.2 60.8 256 63.2 18.0
Macro c ell Indoor Ind oor Urban S 3 111
3840 -108.2 60.8 256 63.2 18.0
User Equipment Transmitter
Maximum UE TX power (dBm) / MEAN UE TX antenna gain (dBi) Total UE TX EIRP (dBm) Base Station Receiver BS RX nois e figure (dB) Bas eline RX Eb/No (dB) (from SMG Mars eilles BS RX Eb/No (dB) BS RX sensitivity (dBm) BS RX antenna gain (dBi) BS RX cable & connector los s es (dB) Maximum Reverse Path Loss Maximum allowable isotropic path loss (dB) DOWNLINK Characteristics
2 1.0 0.0 21.0
2 1 .0 0.0 21.0
3.3 2.2 1.4 -121.5 18.0 3.0
3.3 1.7 0.9 -122.0 18.0 3.0
157.5
158.0
18.0 4.1
18.0 4.1
5.0
5.0
1.6 0.8 -120.4 0.0
1.1 0.3 -120.9 0.0
Base Statio n Transmitter Transmitter
BS TX antenna gain (dBi) BS TX cable & connector los s es (dB) User Us er Equipement Receiv er UE RX nois e figure (dB) Baseline UE RX Eb/No (dB) with Tx Div Bas eline UE RX Eb/No (dB) (from SMG Mars eilles UE RX Eb/No (dB) UE RX RX sens se ns itivity (dBm) UE RX antenna gain (dBi)
14
UMTS site database • Important parameters: - Loca Locattion ion - Ante Antenn nnaa azi azimu muth th - Ante Antenn nnaa down downti tilt lt - Ante Antenn nnaa EIR EIRP P - Ante Antenn nnaa heig height ht - Ante Antenn nnaa typ typee - Pr Pred edic icti tion on mode modell
15
Coverage objective and clutter database
(Barcelona)
16
Terrain database Example
(Barcelona)
17
Design assumptions • Services — Required services: – Speech, – LCD64, 144, 384, 2048 – UDD64, 144, 384, 2048 — Environments - geographical regions – Dense urban – Urban – Sub-urban – Rural
• Service areas of coverage — Accurate geographical geographical definition of each service – Service area polygons 18
Design assumptions • Quality of coverage — Coverage types – Outdoor – Indoor (indoor penetration factor) – In-car or in-train i n-train (in-car or in-train penetration factor) factor) — Quality of coverage – Cell area reliability (Percentage of cell area being covered) Speech
LCD64
UDD64
UDD 144 UDD 384
Urban
Indoor
Indoor
Indoor
Indoor
Indoor
Suburban
Indoor
Indoor
Indoor
Indoor
Indoor
Rural
Indoor
Indoor
Indoor
Outdoor
-
Roads
Incar
-
-
-
19
Design assumptions • Capacities — Per service — Per area — Both Uplink & Downlink — Geographical distribution (traffic maps)
• Traffic models — Busy hour speech traffic :
Erlang
— Busy second data traffic:
Mbit/s
P e a k h o u r tr a ffi c
S p e e ch ( m E rl rl )
Urban Suburban Rural Other
345763 172881 115254 12678
UDD144 Uplink Mbit/s 1,67 0,84 0 0
UDD144 Downlink Mbit/s 10,43 5,22 0 0
UDD384 Uplink (Mbit./s) 9,6 4,8 0 0
UDD384 Downlink (Mbit./s) 19,2 9,6 0 0 20
Design assumptions • All of the design assumptions are correlated — Services — Coverage types
Design targets
— Service areas of coverage — Capacities — Quality of coverage
• Radio network design results are highly dependant of the design assumptions — Any change of one of the assumptions requires to re-do the design work — Design assumptions => Number of sites=>M$ to be invested 21
Lin nk Budge gett Radio dimensioning - Li • Per service • Per morphology — Environment — Coverage type
• Quality of coverage — Cell area reliability
• Capacity & Traffic load — Uplink N-Pole capacity — Uplink interference margin vs. traffic load — Downlink interference margin vs. traffic load — Frequency reuse factor 22
Radio dimensioning - Link Budget • Link budget balance — cell radius based on UL pathloss — determine BS output power per user, assuming: – DL pathloss = UL pathloss
• Performance -Quality of service — BS & MS Eb/No targets per service including diversity gain
• Cell sites — Antenna height, gain — Cable loss — BTS power and Configuration Configuration — Uplink power: mobile TX power 23
Radio dimensioning - Cell count LB @ X% load Design assumptions
Cell size
Cell capacity
# sites for coverage
adjust load
# sites for traffic
Comparison Decision
Final number of sites 24
Radio dimensioning - Multi-service • Multi-services are supported by the same system — LB per service Current Load
ink ink Bud get 1
R adius 1
Radi us 2
in k B u d g e t 2
Radi us n
Cap 1
in k B u d g e t n
Cap 2
Ca p n
25
Radio dimensioning - Paris Example
5 m a c cr r o - o -z o z on e n s e s
26
Radio dimensioning - Paris Example • Service areas and surfaces
Service Area Zone A Zone B Zone C Zone D Zone E
PlaNet polygones Zone1 Zone2 Zone2 Zone3 Zone4
Design criteria * UDD 384 Indoor UDD 384 Indoor UDD 384 Indoor LCD 64 Indoor UDD144 Outdoor
Size
27
Radio dimensioning - Paris Example • A service area is defined by: — Environment (urban, suburban, rural…) — Indoor penetration factor — QOC & QOS
Zone Zone Zone Zone Zone
A B C D E
P é n é t r a tio n 21 dB 18 dB 15 dB 12 dB -
QoC 90% 90% 90% 90% 90% 28
Radio dimensioning - Paris Example
• A service area is defined by: — Required services — Coverage type (indoor, outdoor, in car …)
Zone A Zone B Zone C Zone D Zone E
Speech Indoor Indoor Indoor Indoor Outdoor
UDD 64 Indoor Indoor Indoor Indoor Outdoor
UDD 144 Indoor Indoor Indoor Indoor Outdoor
UDD 384 Indoor Indoor Indoor
LCD 64
LCD 144 Indoor Indoor Indoor
Indoor 29
Radio dimensioning -Paris example 1st Cell Count
2nd Cell Count
Zone 1
Zone 2
F i r rs t s t s e e t t o f f L LB B
S e ec o c n o nd s d e et t o f f L LB B
N1 sites
N2 sites
3rd Cell 3rd Cell Cell Count Count
Zone 3
t h h i ir r d d s e et t o f f L LB B
N3 sites
4th Cell 4th Cell Cell Count Count
5th Cell 5th Cell Cell Count Count
Zone 4
f o u ur r t t h h s e et t o f f L LB B
N4 sites
Zone 5
f i f f t h h s e et t o f f L LB B
N5 sites
Total # of sites
Number of sites is evaluated on the basis of a coverage and a traffic study 30
Radio planning - Planning tool • W-CDMA technology • Multi-service — Speech — Data: – LCD64, 144, 384, 2048 – UDD64, 144, 384, 2048
UMTS Cell Quality Coverage Planning of Maps ToolService
UMT MTS S Ce Cell ll Planning Tool
• Multi-carriers • Different user types
31
Pred edict iction ion & Simu Simula latio tion n Radio planning - Pr BTS
MS
PA
link losses Noise figure Eb/No Processing Gain ...
Site Cable
PA
Tx/Rx
Noise
Service
figure
Eb/No Processing
Gain
...
loss Antenna pattern sites configuration sites constraints (reuse of existing sites )
Terrain Terrain Data Data Base Base Propagation Propagation Model Model Monte-Carlo Monte-Carlo Simulations Simulations
Statistics dropped calls hand-offs UL noise rise Cell loading…
Maps sites Location & configuration
Pilot Ec/Io UL Requ Require ired d MS E EIRP IRP Soft HO 32
Radio planning - Planning tool Cell Planning Tool
• Coverage prediction — Pathloss calculations — Coverage based on design thresholds Coverage Maps
Quality uality of Service
• QOS simulat simulations ions-- Monte Monte Carlo simulat simulations ions — UL/DL Eb/N0 targets — Power allocation — Power control — Soft HO
Call drops Uplink noise rise HO regions and statistics
33
Radio planning Coverage analysis
User input BS radiated power
Height database
Clutter database
antenna pattern frequency
Received power at each bin
• Design Threshold — based on maximum allowable path loss for the most constraining service. — The required mobile received signal is calculated based on BS EIRP and on Maximum Path Loss. Design threshold=BS EIRP-Maximum Path Loss 34
Radio planning -Monte Carlo simulations
•Users are randomly generated based on • Poisson law •Traffic maps •Successive shots 35
Radio planning - Pilot Ec/Io
36
Radio planning - UL Required Mobile EIRP
37
Radio planning - RF Design Optimisation
optimisation: • Standard optimisation: — Antenna re-orientation — antenna downtilt — site re-location
• Site densification — increase capacity
Continuos process
— increase the level of interference
• carrier overlay
38
Conclusions • Optimum network radio design requires accurate design assumptions in terms of services, coverage, capacity, and quality of service
• Capacity and coverage strongly correlated
• Multi service — voice and data over the same radio resource 39
