LTE RF Design and Optimization

LTE RF Design and Optimization

Optimi Operator’s Workshop Oct. 6th & 7th, 2009

1

CUSTOMER CONFIDENTIAL

LTE RF Design and Optimization Layout

Introduction Overall RF Design and Optimization Process Design Input Design Objectives Pathloss Model Benchmarking of LTE Design Objectives Site Selection and RF Optimization Summary

2


Introduction UTRAN Long Term Evolution (LTE)

LTE belongs to the next generation of mobile systems recently standardized in 3GPP Orthogonal Frequency Division Multiplexing (OFDM) Adaptive modulation and coding with hybrid ARQ Fast packet scheduling with full flexibility in time and frequency Full spectrum flexibility with BW ranging from 1.4 to 20 MHz Standardized MIMO support with up to 4 antennas on each side

3


Overall RF Design and Optimization Process

Fine Tuning

General Overview

4

RF Planning Tool Predictions

Path-loss predictions are obtained based on propagation models maybe combined with drive tests and OSS

Static Simulator Nominal Design

SINR, data rate and quality network information is provided based on pathloss [more details in next slides]

Optimizer Tuned Configuration

Sites are selected and tilt, azimuth, power, etc. are tuned to improve the performance within the specified constraints

Static Simulator Improved Performance

Results are analyzed via a static simulator. The process can be repeated for a finer tuning

RF Planning Tool Re-assessing Predict.

After optimization predictions may be re-calculated for the sake of providing better accuracy

Static Simulator Optimized Performance

Optimized results are obtained using re-calculated predictions from the RF planning tool


Design Input Basic Parameters

Physical Parameters Terrain attributes, clutter type, antenna location (latitude and longitude) and antenna configuration (azimuth and tilts)

Generation of Predictions Propagation models, drive tests, OSS data, call traces, etc.

eNode-B Parameters: PA power, pilot power, cyclic prefix, IoT level, network load, noise figure, etc.

UE Parameters TX power, antenna gain, noise figure, etc.

Duplexing Mode Frequency Division Duplex (FDD) different channels for DL and UL Time Division Duplex (TDD) sharing in time a single frequency for DL and UL

5


Design Input Link Level Mapping Table 7 MCS-1 [QPSK,R=1/8] MCS-2 [QPSK,R=1/5]

Throughput, bits per second per Hz

6

MCS-3 [QPSK,R=1/4] MCS-4 [QPSK,R=1/3]

5

MCS-5 [QPSK,R=1/2] MCS-6 [QPSK,R=2/3] MCS-7 [QPSK,R=4/5]

4

MCS-8 [16 QAM,R=1/2] MCS-9 [16 QAM,R=2/3]

3

MCS-10 [16 QAM,R=4/5] MCS-11 [64 QAM,R=2/3] MCS-12 [64 QAM,R=3/4]

2

MCS-13 [64 QAM,R=4/5] Shannon

1

0 -10

-8

-6

-4

-2

0

2

4

6

8

SNR, dB

3GPP TR 36.942 V8.2.0 6


10

12

14

16

18

20

Design Input Standard Propagation Models Lee Model Empirically derived area model that is commonly used in the United States Wireless applications in the 800MHz and 1900MHz range. Applied at higher frequencies, but adjustments must be made to the slope and intercept

Hata Model Most-popular empirically-derived propagation model for the 800MHz to 2GHz frequencies Widely used in Asia accurately describing the dense urban environments better Based on the Japanese propagation environment, different from USA or Europe areas.

COST 231 Model Up-banded version of Hata Model adjusted for 1800-1900MHz frequency band. Enriched with correction terms: street width, orientation, building height, etc. Flexible model frequently used both as a macroscopic and a microcell model.

SUI Model An extension of the earlier work by AT&T Wireless and Erceg et al. Widely used for technologies at frequency band higher than 2GHz Selected to test WiMAX due to its accurate estimations at NLOS environments

7


Design Input Multiple Input Multiple Output (MIMO)

TX/RX Diversity

eNode-B …

…

…

UE SINR Gain

Spatial Multiplexing

eNode-B …

…

…

UE Slight decrease in SINR Large Throughput Boost

Feedback (Closed-Loop) – Better SINR 8


Design Input Multiple Input Multiple Output (MIMO) (II)

MIMO capability is key feature in LTE to achieve Ambitious requirements for throughput High spectral efficiency

Receive and/or Transmit Diversity Same information is sent/received over multiple antennas Gain in SINR

Open-Loop Spatial Multiplexing Different information is sent/received over multiple antennas Decrease in SINR due to higher interference but large boost in throughput

Closed-Loop Spatial Multiplexing Same approach as before, but getting advantage of feedback information Improves de SINR at a cost of more complexity

9


Design Input Uplink Power Control (PC)

Classic PC schemes aim all users received with the same SINR 3GPP agreed the use of Fractional PC for Physical Uplink Shared Channel (PUSCH) to compensate for slow channel variations

{

PTX = min Pmax , P0 ⋅ N RB (n) ⋅ Lα

}

Pmax is the maximum user transmit power P0 is a sector-specific parameter NRB is the number of allocated RBs L is the downlink pathloss α is the pathloss compensation factor

Users with higher pathloss operate at lower SINR requirements Interference to neighbors decrease Overall system performance tend to improve

10


Design Input Resource Block Planning (I)

Resource Block (RB) planning is a key factor on interference control A smart allocation can significantly improve the system performance

The radio access technology may impact the RB planning strategy OFDMA (in DL) allows allocation of non-contiguous bandwidth SC-FDMA (in UL) forces to allocate contiguous bandwidth

Traditional RB schemes Full reuse: all sectors within a site share the same bandwidth Higher peak throughput at a cost of higher interference

One-third reuse: bandwidth shared among the sectors within a site Lower peak throughput but getting an improvement on SINR

Advanced RB schemes Dynamic RB Planning: automatic solution to minimize the interference Inter-Cell Interference Coordination (ICIC): wiser allocation scheme in between full and one-third reuse

11


Design Input Resource Block Planning (II) - ICIC

Wise allocation of users generating higher interference to improve the system performance Cell-edge users, which are assumed to interfere the most, have a limited band to be scheduled Rest of the bandwidth for cell-center users

Interference coordination Cell-edge band location follows the well-known 3-color pattern within a site Distance between highly interfering users increases cell-edge

cell-center

System bandwidth

12


Design Input Resource Block Planning (III) - ICIC

No ICIC

ICIC with 3 dB offset

Worse SINR in CC

Better SINR in CE

13


Design Objectives

Resource Block (RB)

Carrier Bandwidth

LTE Metrics

RSRP Average RX power of one RE transmitting RS

RSRQ RSRP x # RBs Carrier RX power + Noise Time Slot

Resource Element (RE)

14


Design Objectives Coverage (I)

Indicating if a certain location may have access to the network Defined by Reference Signal Received Power (RSRP) Linear average over the power contributions of the REs that carry cell-specific Reference Signals (RSs) within the considered frequency bandwidth

15


Design Objectives Coverage (II) – Radio Link Budget [Downlink] a

Transmit Power [dBm]

43.0

b

TX Antenna Gain [dBi]

13.0

c

Cable Loss [dB]

d

EIRP [dBm]

e

UE Noise Figure [dB]

5.0

f

Thermal Noise [dBm]

-106.8

= k (Boltzmann) x T (300K) x B (5MHz)

g

Received Noise Floor [dBm]

-101.8

=e+f

h

SINR [dB]

i

Receiver Sensitivity [dBm]

j

Interference Margin [dB]

3.0

k

Control Channel Overhead [dB]

1.0

l

RX Antenna Gain [dBi]

0.0

m

RX/TX Diversity Gain

3.0

N

Body Loss [dB]

0.0

Maximum Pathloss [dB [dB]] 16

5 MHz – MIMO 1x2 – 10 Mbps

0.0 56.0

10.0 -91.8

146.8

=a+b–c

for 16QAM 2/3 =g+h

=d–i–j–k+l+m–n


Design Objectives Coverage (III) – Radio Link Budget [Uplink] a

Max Transmit Power [dBm]

b

TX Antenna Gain [dBi]

0.0

c

Body Loss [dB]

0.0

d

EIRP [dBm]

24.0

e

eNode-B Noise Figure [dB]

10.0

f

Thermal Noise [dBm]

g

Received Noise Floor [dBm]

h

SINR [dB]

i

Receiver Sensitivity [dBm]

j

Interference Margin [dB]

2.0

k

Cable Loss [dB]

0.0

l

RX Antenna Gain [dBi]

m

RX/TX Diversity Gain

0.0

n

MHA Gain

2.0

Maximum Pathloss [dB [dB]] 17

24.0

-106.8 -96.8 6.0 -90.8

5 MHz – SISO 1x1 – 5 Mbps =1+2–3

= k (Boltzmann) x T (300K) x B (5MHz) =e+f for QPSK 2/3 =g+h

13.0

127.8

=d–i–j–k+l+m+n


Design Objectives Quality

Giving an idea of the level of interference, highly impacting the performance Defined by Reference Signal Received Power (RSRQ) RSRP over the wideband received signals from all base stations in the carrier bandwidth plus thermal noise

18


Design Objectives Capacity

Traffic maps represents Active subscriber’s population spatial distribution Overall offered load that need to be served by the network

Demand grid, i.e. user spatial location, is based on clutter types Active users in a dense urban area is much higher than in forest areas Accuracy improves by network measurements from active users

Marketing information defines traffic volumes and service mixes So that it is possible to derive the network offered load

Note that each service has specific requirements and hence need to be assigned to different radio access bearers (RAB) Requested Data Rate: throughput for a user to be satisfied Minimum Data Rate: throughput for a user to be in the system

19


Pathloss Model RF Planning Tool

Basic

Purely Predictions

Interpolation and Drive Tests

Very vulnerable to database errors and prediction inaccuracy

Extra accuracy and robustness against database errors

OSS Based

No need for Drive Test. Extra accuracy from relaying on OSS data

Geolocation

Enhanced accuracy from geolocated events

Advanced

20


Benchmarking of LTE Design Objectives Flow Diagram Pathloss

Project Build Sector

Service

User

System

Configuration

Monte Carlo Simulator

Neighbor List

Reports

Analysis Coverage SINR Data Rate Quality ...

21


Benchmarking of LTE Design Objectives Monte Carlo (MC) Simulator

Monte Carlo simulation solution is used to characterize the radio performance of LTE at any time of the design process Quick identification of the best design among multiple candidates Clearly pointing the main network problems (highest blocked/dropped field)

Required inputs System, sector and user parameters Service and traffic set-up

Provided outputs Accurate estimations for UL and DL loading, and noise rise Different raster views and text-formatted reports with information about served, unsatisfied and drop users, offered and carried loading, etc.

22


Benchmarking of LTE Design Objectives Reason for Failure

23


Benchmarking of LTE Design Objectives Downlink Data Rate

24


Site Selection and RF Optimization Definition of Network Planning Criteria

KPIs per clutter RSRP threshold: minimum RSRP level to consider a pixel as covered. RSRQ threshold: minimum RSRQ level to consider a pixel as in good quality. Weight in order to differentiate the relevance of a clutter type. Penetration Loss in order to add extra losses.

Global KPIs Coverage: percentage of covered area from signal level (RSRP) perspective. Quality: percentage of covered area from quality level (RSRQ) perspective Traffic Quality: percentage of covered area for minimum SNR based on minimum data rate the service requires Capacity: percentage of sectors at maximum load

Financial cost component Monetary cost per RF change and sector Necessary to make sure that the proposed RF design meets the budgetary constraints

25


Site Selection and RF Optimization Accomplishing KPI Objectives

The location of potential site comes from An existing network, e.g. UMTS. A “random” deployment over the area of study

KPIs fullfiled with just 60% of initial locations

Site selection minimum number of sites to meet target coverage, quality and capacity. 15 sites

26

KPI performance from potential sites

KPI Objectives


KPI performance after site selection

9 sites

Site Selection and RF Optimization Selected Sites

27


Site Selection and RF Optimization Optimizing Antenna Setting (I)

Operators have limited amount of resources, but at the same time they require to fulfill certain Key Performance Indicators (KPIs). The optimization process aims to improve the overall network coverage, capacity and quality, and enabling operators to make the most out of their limited network resources. Network attributes that can be modified: Antenna type Antenna height Antenna tilt (mechanical and electrical) Antenna azimuth Transmit power

28


Site Selection and RF Optimization Optimizing Antenna Setting (II)

RSRQ Coverage - 81.87% to 85.11%

29


Site Selection and RF Optimization Combined Solution

KPIs are fulfilled with 1 site less

30


Other Functionalities Cell-ID Planning

According to 3GPP there are 504 unique physical-layer cell identities The different cell-IDs are grouped into 168 unique physical-layer cell-ID groups Each group containing three unique identities Each cell-ID is part of one and only one physical-layer cell-ID group

Cell-ID planning aims to Maximize the radio distance between cell-IDs Avoid (or minimize) the amount of neighbors with the same cell-ID

31


Summary Conclusions and Remarks

LTE is a new technology recently standardized by 3GPP Network deployment is still in study phase Operators can clearly benefit from Efficient site selection (based on current 3G sites) Optimized antenna configuration to maximize performance More accurate pathloss models

LTE key metrics for optimization RSRP to indicate the network access (i.e. coverage) RSRQ giving an idea of the link quality Capacity which is determined by traffic and distribution of users Advanced LTE features also have an impact on the design MIMO capabilities to improve SINR and throughput RB planning (and ICIC) to control interference Smart schedulers to optimize RB allocation

32


THANK YOU! Comments and Questions?

33


LTE RF Design and Optimization

Recommend Documents