ChEG
Updated Calibration Information of Chinese Evaluation Group
1
Introduction
On ITU-R WP5D #5 meeting in June of 2009, Chinese Evaluation Group (ChEG) submitted
document “Calibration Activities in Chinese Evaluation Group” to provide calibration information
and approach performed in ChEG.
This document will provide more detailed approaches of calibration. And newest results will also be
provided.
2
Calibration activities in ChEG
For calibration among its members, a step-by-step approach has been taken by ChEG. Following
steps have been defined and results are included in next section for reference.
3.1
Channel model calibration
The intention of this step is to calibrate the implementation of newly defined IMT-Advanced
MIMO channel. The output is in form of statistical characteristics of MIMO fast fading..
3.2
1)
2)
3)
4)
3.3
Link level calibration
Link performance in AWGN channel
This step is to calibrate the implementation of basic transceiver algorithms among ChEG
members. The output is in form of link level curve.
Link performance in MIMO fading channel (with ideal channel estimation)
This step is to calibrate the implementation of MIMO transceiver algorithms among ChEG
members. The output is in form of link level curve.
Link performance in MIMO fading channel (with real channel estimation)
This step is to calibrate the implementation of channel estimation algorithms among ChEG
members. The output is in form of link level curve.
Link performance in MIMO fading channel with adaptive scheduling and HARQ
This step is to calibrate the implementation of full link level processes, including channel
measurement / feedback, Adaptive Scheduling, AMC, HARQ, etc. The output is in form of
link level spectrum efficiency.
System level calibration
Here ChEG referred the methodologies taken by 3GPP in LTE-advanced self evaluation. Those
include:
1)
Large scale fading characteristics of system level platform.
This step is to calibrate the basic characteristics of system level platform among ChEG
members. Those characteristics include distribution of downlink wideband SINR (also
known as geometry) as well as coupling loss.
2)
System performance in simple scenario
This step is to calibrate the system level performance with simple system level simulation
assumption, e.g. SIMO, Round Robin scheduling, etc. Output is in form of the distribution
of user SINR and throughput.
ChEG
4
Approach and result of calibration in ChEG
In this section, we provide calibration approaches and results which are the latest outcomes of
calibration activities of ChEG.
4.1
•
Channel model calibration for fast fading
Simulation approach
Channel model calibration is performed to calibration fast fading parameters, i.e. delay, AoD and
AoA. Statistical characteristic of IMT-Advanced MIMO fast fading parameters, RMS is used here.
To avoid multiple uncertain factors are calibrated in one process, the calibration is divided to:
 LOS case
 NLOS case
 final calibration
In calibration of LOS case, calculation and statistic are done only for LoS users. So LoS users in
system (57 sectors) shall be selected or all users in system can be set as LoS users. For one user,
parameter (power, delay, AoA, AoD) in RMS formula is in unit of cluster (RMS calculation method
can be found in 25.996 ANNEX A). The parameter values are obtained before cluster divided to
sub-cluster in system level platform. And normalized power Pn is obtained after
formula Pn 
Pn
1
K R 1

N
P
n1 n
 n 1P1, LOS . So RMS curves of delay, AoD and AoD can be
achieved. Ideal curve according the distribution of delay, AoA and AoD can also be drawn.
Calibration for NLoS case is similar to LoS case, expect that only NLoS users are selected or all
users are set as NLoS users.
For final fast fading calibration including LoS and NLoS, all users including NLoS ones and LoS
ones as specified in M.2135 are counted. The statistic approach is similar to LoS and NLoS case,
except that parameter in RMS formula is in unit of ray for each user.
•
Result
Calibration results on statistical characteristic of IMT-Advanced MIMO fast fading channel are
shown in the following embedded documents. Result for all mandatory scenarios (indoor, UMa,
UMi and RMa) are included.
Final
LoS of
NLoS of
IMT-Advanced fast IMT-Advanced
fading calibration_ChEG
fast IMT-Advanced
fading calibration
fast fading calibration
ChEG
1)
•
•
Link level calibration
Link performance in AWGN channel
Simulation assumptions
– Antenna configuration: 1x1;
– Turbo decoding algorithm: Max-Log-MAP with 8 iterations;
– Others:
Short code length
Middle code length
Long code length
CQI
5
9
15
Modulation
2
4
6
TBS
392
1128
3112
rate coding
0.4385
0.601
0.9258
SNR
[−1 : 0.5 : 2]
[7.8 : 0.2 : 9]
[19.4 : 0.2 : 20.6]
Result calibration
1
Lp=392 comp1
Lp=1128 comp1
Lp=3112 comp1
Lp=392 comp2
Lp=1128 comp2
Lp=3112 comp2
Lp=392 comp3
Lp=1128 comp3
Lp=3112 comp3
Lp=392 comp4
Lp=1128 comp4
Lp=3112 comp4
Lp=392 comp5
Lp=1128 comp5
Lp=3112 comp5
Lp=392 comp6
Lp=1128 comp6
Lp=3112 comp6
0.1
0.01
1E-3
BER
1E-4
1E-5
1E-6
1
1
2
3
0.1
4
5
0.01
BLER
4.2
6
1E-3
1
A
Lp=392 comp1
Lp=1128 comp1
Lp=3112 comp1
Lp=392 comp2
Lp=1128 comp2
Lp=3112 comp3
Lp=392 comp4
Lp=1128 comp4
Lp=3112 comp4
Lp=392 comp5
Lp=1128 comp5
Lp=3112 comp5
Lp=392 comp6
Lp=1128 comp6
Lp=3112 comp6
Lp=392 comp7
Lp=1128 comp7
Lp=3112 comp7
A
B
C
D
E
F
G
1E-7
7
1E-4
1E-8
1E-9
-2
0
2
4
6
8
10
12
14
16
18
1E-5
20
-2
0
2
4
SNR
2)
•
6
8
10
12
14
16
18
20
SNR
Link performance in MIMO fading channel (with ideal channel estimation)
Simulation assumptions
Parameter
Value
Scenarios
UMA NLoS, 30 km/h (CDL)
Transmission scheme
4*2
Precoding vector: [1,0;1,0;0,1;0,1]
Channel estimation
Ideal
RB/block size
4RB Pair (1 ms)/(1 128 + 24)*2 codeword)/16QAM
simulation time
10000 packages
Antenna configuration in base station
10 lambda spacing
Antenna configuration in UE
0.5 lambda spacing
MIMO receiver
MMSE
ChEG
•
Result calibration
3)
•
Link performance in MIMO fading channel (with real channel estimation)
Simulation assumption
Channel estimation algorithm
•
Result calibration
Note：s1 means stream 1, s2 means stream 2.
2D Viena filter/LMMSE with linear interpolation
ChEG
4)
•
Link performance in MIMO fading channel with adaptive scheduling and HARQ
Simulation assumption
Parameter
Value
Scenarios
UMA NLoS, 30 km/h
Channel model
CDL
Antenna configuration
1*4, 10 lambda spacing
Channel estimation
Real
RB size
4RB Pair (1 ms)
HARQ Timing
Refer to LTE R8 uplink
SNR
4.8 dB
Output
Spectrum Efficiency
AMC
On, refer to LTE R8 uplink
•
Result calibration
Output Spectrum Efficiency from members: 1.50~1.56 bps/Hz
4.3
1)
•
System level calibration
Large-scale fading characteristic of system level platform
Simulation assumption
•
Parameter
Value
Handover margin
Shadow fading and LOS correlation between sites for ITU scenarios
Shadow fading and LOS correlation between sectors of the same site for ITU
scenarios
BS feeder loss
BS antenna downtilt in indoor hotspot scenario (InH)
BS antenna downtilt in urban micro-cell scenario (UMi)
BS antenna downtilt in urban macro-cell scenario (UMa)
BS antenna downtilt in rural macro-cell scenario (RMa)
LOS pathloss propagation for indoor users in UMi
User distribution
1dB
0
1
2dB for ITU scenarios
0 deg
12 deg
12 deg
6 deg
Yes
Re-drop within minimum
distance
Result
Distributions of downlink wideband SINR as well as coupling loss for four mandatory
scenarios are included in embedded documents.
Large-scale
fading characteristic of system level
2)
•
System performance in simple scenario
Simulation assumption
Parameter
Value
General
Parameters and assumptions not explicitly stated
here according to ITU guidelines M.2135 and 3GPP
specifications
ChEG
Duplex method
FDD
Network synchronization
Synchronized
Handover margin
1dB
Downlink transmission scheme
1x2 SIMO
Downlink scheduler
Round robin with full bandwidth allocation
Downlink link adaptation
Wideband CQI, no PMI on PUCCH (mode 1-0)
5ms periodicity,
6ms delay total (measurement in subframe n is used
in subframe n+6)
CQI measurement error: None
MCSs based on LTE transport formats [36.213]
Downlink HARQ
Maximum four transmissions,
Downlink receiver type
MRC
Uplink transmission scheme
1x2 SIMO
Uplink scheduler
Frequency Domain Multiplexing – non-channel
dependent, share available bandwidth between users
connected to the cell, all users get resources in every
uplink subframe.
With M users and Nrb PRBs available,
Mh=mod(Nrb,M) users get floor(Nrb/M)+1 PRBs
whereas Ml=M-Mh users get floor(Nrb/M) PRBs
Uplink Power control
P0 = -106dBm, alpha = 1.0
Uplink Link adaptation
Based on delayed measurements. Ideal channel
estimate from UL transmission in subframe n can be
used for rate adaptation in subframe n+7
MCSs based on LTE transport formats [36.213]
Uplink HARQ
Maximum four transmissions
Proponent to specify IR or CC
Uplink receiver type
MMSE in frequency domain,
MRC over antennas
(no intercell interference rejection)
Antenna configuration
Vertically polarized antennas
0.5 wavelength separation at UE,
10 wavelength separation at base station
Channel estimation
Ideal, both demodulation and sounding
Control Channel overhead,
Acknowledgements etc.
LTE: L=3 symbols for DL CCHs, M=4 resource
blocks for UL CCH, overhead for demodulation
reference signals,
BS antenna downtilt
InH: N/A
Umi: 12deg
Uma: 12deg
Rma: 6 deg
Feeder loss
0dB
Channel model
According to ITU for ITU scenarios
Intercell interference modeling
Explicit
ChEG
•
Results
Distributions of user SINR and throughput for four mandatory scenarios are included in
embedded documents.
distribution of
user SINR
distribution of
user throughput