LTE System Level Performance in the Presence of

CQI Feedback Uplink Delay and Mobility

Kamran Arshad
Mobile and Wireless Communications Research Laboratory
Department of Engineering Systems
University of Greenwich, United Kingdom
Email: k.arshad@greenwich.ac.uk

A BSTRACT static BLER curves under a specific channel model can be

Long Term Evolution (LTE) is a promising technology for generated either by link level simulations or mathematical
the wireless and mobile communication systems and is being models [3]. Finally, the selected values are sent to the eNodeB
commercially launched in various countries of the world. The as a CQI feedback vector. Based on the received CQI, eNodeB
Third Generation of Partnership Project (3GPP), through de- perform resource scheduling and MCS selection.
sign and optimisation of new radio access techniques, is further Packet level scheduling is one of the core functionality of
developing the future LTE-Advanced technology to remain LTE and plays an important role in the optimisation of the
on the forefront of wireless technologies. However, it is still network. A scheduling scheme can be designed to allocate
challenging in LTE to provide the desired quality of service each UE a portion of the available resources i.e. resource
to the users particularly in the presence of high mobility and blocks (RBs). For an efficient scheduling, ideally each UE
uplink feedback delay. In this paper, the impact of Channel needs to send periodically a set (one for each sub-band)
Quality Indicator (CQI) uplink feedback delay on the overall of CQI values to the eNodeB. However, this would result
network performance under different scheduling algorithms a tremendous uplink control signalling overhead. Further, in
and mobility patterns is discussed. We study average user practice, delays of several transmission time intervals (TTIs) in
equipment (UE) throughput, average cell-edge UE throughput the reception of CQI feedback at eNodeB may exist that causes
and average cell throughput under mobility and conclude that severe performance degradation. In fast fading channels, the
for an efficient LTE-Advanced scheduling algorithm, UE speed CQI at the scheduling instant greatly deviates the instant when
and CQI feedback delay must be taken into account. CQI was generated and hence negatively affect the accuracy
of MCS selection [4].
I. I NTRODUCTION Several CQI feedback reporting and compression schemes
LTE was standardised by 3GPP, initially in release 8 have been proposed in the literature [5]. The degradation of
(December 2008), now commercially available in various system performance due to the delayed CQI feedback was
countries all around the world. LTE specifications are now analysed in [6], [7]. Depending on the scheduling scheme,
stable and several enhancements were added in subsequent an eNodeB may select the UE with highest sub-channel CQI
3GPP releases (i.e. R9 - R13). LTE and LTE-Advanced values to maximise the system throughput. As 3GPP did not
specifications are targeting to achieve higher peak data rates, specify scheduling mechanisms for both uplink and downlink,
higher spectral efficiency, lower latency and all IP based a large number of scheduling algorithms have been proposed
optimised network [1]. Higher data rate, and spectral efficiency in the literature, for more details reader may refer to [8].
is achieved by using flexible, scalable carrier bandwidths There are several technical challenges that will need to
ranging from 1.4MHz up to 100MHz, Orthogonal Frequency be addressed in the future releases of 3GPP LTE (i.e. LTE-
Division Multiple Access (OFDMA) in downlink, Adaptive Advanced). In this paper, we show that the impact of CQI
Modulation and Coding Scheme (A-MCS) and Multiple Input feedback delay becomes worst as UE speed increases and
Multiple Output (MIMO). hence design of a scheduling algorithm incorporating UE
In order to achieve target data rate or throughput, eNodeB speed for future LTE releases is an important task. The aim
selects the best MCS for each sub-carrier depending on the of this paper is to investigate the impact of UE speed on CQI
channel state information (CSI). At the UE, a module is feedback delay under different scheduling algorithms proposed
responsible for processing the CSI information, for instance, in the literature, more specifically, we consider best cqi (Bcqi),
estimating sub-carriers received Signal to Interference Noise round robin (RR) and proportional fair (PF) algorithms [8].
Ratio (SINR) to obtain a Block Error Rate (BLER) estimate The simulations were performed using the LTE system level
under the assumption of a particular channel model. An offline simulator discussed in [2]. The outcome of this study will
lookup table can be used to map this information into a higher provide a guideline for the design of scheduling algorithms
rate CQI value that can specify a 10% BLER target [2]. Such for the future LTE-Advanced systems.
978-1-4799-6532-8/15/$31.00 c 2015 IEEE The remainder of this paper is organised as follows: sec-
tion II describes the preliminaries: overview of LTE downlink, The exponential effective signal to interference and noise
CQI measurements and feedback. This is followed by a brief ratio mapping (EESM) is used in the simulations to obtain
overview of LTE schedulers in section III. Simulation details, the eff , which can be used to map the TB BLER from either
results and discussion are provided in section IV. Finally, mathematical models or link level simulations.
section V concludes this paper. ✓ ◆!
1 X k
eff = · log exp (3)
II. S YSTEM M ODEL N
k2N
A. LTE Downlink Overview
where N is the number of sub-carriers and is a calibration
The 3GPP LTE standard supports both frequency division parameter to fit the compression function in (3) to AWGN
duplex (FDD) and time division duplex (TDD) and can be BLER curves. The eff is finally mapped by means of an
deployed over a wider range of bandwidth (1.4MHz, 3MHz, AWGN BLER curve of the corresponding MCS to a BLER
5MHz, 10MHz and 20MHz). In this paper, we consider only value. Fig. 1 shows the BLER curve for each CQI value
FDD with 20MHz bandwidth for the sake of simplicity and in AWGN channel and each CQI value represents a unique
similar conclusions can be drawn for other bandwidth as well. combination of modulation scheme and coding rate. From
In OFDMA based LTE downlink, the minimum resource that Fig. 1, the CQI feedback value can be obtained using a simple
can be assigned to a user is referred as a RB. A RB is defined step function, Q = f ( eff, dB ) as shown in Fig. 2.
as a group of 12 sub-carriers (or 180KHz) in the frequency
domain and a half TTI i.e. 0.5ms slot in the time domain (1 0
LTE BLER for CQIs 1 to 15
10
TTI = 2 ⇥ 0.5ms slot). In each TTI, a radio frequency carrier
consists of any number of RBs ranging from a minimum of
6 to a maximum of 110 RB that corresponds transmission
bandwidth from 1.4MHz to 20MHz with a granularity of
15kHz [3]. In LTE, scheduling is done on a TTI basis, hence
−1
10

a RB is assigned to a user for two consecutive time slots. In

a time slot, 7 OFDM symbols are transmitted hence if 1 RB
BLER

is assigned to a user for one TTI, a total of 168 OFDM sub-

carriers transmits. The physical layer (PHY) receives user data −2
10

from the MAC layer in the form of an entity called transport

block (TB).
B. CQI Measurements
−3
UE makes measurement of SINR at each RB and then 10
−10 −5 0 5 10 15 20 25

convert the estimated value of SINR into a corresponding CQI

SINR [dB]

for which the received TB BLER shall not exceed 10% [3]. Fig. 1. BLER curves for CQI values from CQI 1 (leftmost, dash) to
In this paper, we assume a multi-cell multi-user (MC-MU) CQI 15 (rightmost, dash-dot) for 1.8GHz carrier frequency, 20MHz channel
scenario where a cell is surrounded by a number of interfering bandwidth and 1000 TTI simulation time
cells, say 'int . The received SINR at k th sub-carrier can be
written as C. CQI Feedback
|h0,k |2
L0,k P0,k According to LTE standard, each UE measures downlink
k = P' 2 (1)
int |hi,k | 2
i=1 Li,k Pi,k + ⇣w
channel quality in terms of estimated SINR. The SINR-to-CQI
mapping is realised and an appropriate CQI value is selected
where P0 is the received power from UE own eNodeB, Pi is as explained in section II-B. The obtained CQI values, then,
the received power from ith interfering eNodeB, L is the path reported back to eNodeB using a CQI reporting scheme on a
loss, h is the complex channel gain and ⇣w2
is the noise power. particular control channel (HS-DPCCH) which may arrive at
We assume an urban environment based on 3GPP TS 36.942 eNode with a certain delay. The feedback delay time comprises
path loss model [9] and ITU Vehicular-B channel model is of processing time, transmission time, and time taken when
used for the generation of channel coefficients. waiting for a scheduling slot [5]. The delay in the reception
Due to the channel frequency selectivity, received SINR on of CQI feedback at eNodeB may cause significant performance
RBs differs from each other. In order to estimate CQI, SINRs degradation. Depending on how quickly scheduler assigns
on different RBs must be mapped into an effective SINR, radio resources to the users, the delay may be about 1 to 4
obtained from mapping the sub-carrier SINR to an AWNG TTI (each TTI is 1ms).
equivalent SINR. The BLER of the effective SINR, eff in
AWGN channel should match the original TB BLER as III. OVERVIEW OF LTE S CHEDULERS
Scheduling is a process of dynamically allocating RBs
pe ( eff ) = pe ( k, k 2 C {b}) (2)
among UE based on scheduling algorithm and the received
where C {b} is the set of RBs in sub-band b. CQI values. The scheduler is located at an eNodeB and decides
 16
SINR−CQI mapping function TABLE I
S IMULATION PARAMETERS
15

14
Simulation Parameter Value
13
Carrier frequency 1.8 GHz
12
Bandwidth 20 MHz
11

10
Simulation scenario SISO, Macro cell
9
Simulation time 1000 TTI
CQI Value

8 TTI length 1 ms
7 Number of cells 19 tri-sector eNodeBs
6 UE per eNodeB per sector 100
5 Path loss model TS 36.942, urban [9]
4 Std. deviation of shadow fading 10dB
3 eNodeB Tx Power 40dB
2
SINR averaging EESM [2]
1
FFT size 2048
0
−20 −15 −10 −5 0 5 10 15 20 25 30 Data sub-carriers 1200
SINR [dB]
Cyclic Prefix Normal [3]
Sampling frequency 30.72 MHz
Fig. 2. SINR to CQI mapping function consisting of 10% BLER points
from Fig. 1 Uplink feedback delay {0, 1, 2, 3, 4} TTI
UE speed {5, 30, 70, 120} Km/Hr

which users to be scheduled and which RBs need to be

allocated in a given TTI by taking into account various factors to the UE with relatively higher CQI i.e. at each time instant
such as throughput, spectral efficiency, fairness etc. [8]. the k th UE is selected as
Ri
A. best cqi (Bcqi) scheduler k = arg max (5)
i Ri
In this scheduling strategy, the radio resources are assigned where Ri is the average data rate of the ith UE.
to the users with best channel conditions (alternatively, users
with higher values of CQI). In Bcqi the channel selected for IV. S IMULATION R ESULTS AND D ISCUSSION
transmission have high quality and with appropriate MCS, a We assume an OFDMA based downlink MC-MU scenario,
higher throughput or higher system capacity can be achieved. consisting of 19 cells with identical 3 sector eNodeB’s. The
The Bcqi increases cell capacity at the expense of fairness, UE are uniformly distributed within each eNodeB sector. The
as UEs with bad channel conditions (e.g. cell edge UE) are wireless link attenuation consists of (1) a deterministic path
unlikely to be scheduled. Mathematically, loss urban model (2) a zero mean log-normally distributed
shadow fading, and (3) a small scale Rayleigh fading. In all
k = arg max Ri (4)
i simulations, we assume ITU Vehicular-B power delay profile,
as it is more realistic when users are moving at higher speeds.
where Ri is the instantaneous data rate of ith user. We consider a full buffer traffic model in our simulations
that is there is always data available for every user. Other
B. Round Robin (RR) scheduler
simulation parameters are summarised in Table I.
In this scheduling strategy, UE’s are assigned RBs in In this paper, system level performance is evaluated in
turn one after another without taking into consideration the terms of the average UE throughput, average cell-edge UE
instantaneous channel conditions (or CQI values). Round robin throughput and average cell throughput under the following
scheduling is a fair scheduling scheme in a sense that same scenarios: very low mobility (5 Km/Hr), average mobility
amount of radio resources (i.e. RB) are given to all users. Cell (30 Km/Hr), high mobility (70 Km/Hr) and very high mo-
edge users with worst channel conditions get an equal share of bility (120 Km/Hr). For each scenario, we consider all three
radio resources, however, RR is not fair in a sense of providing schedulers explained in section III. Average UE throughput
the same quality to all users. As RR scheduling does not take for Bcqi, RR and PF scheduler for different CQI feedback
into account the instantaneous channel conditions or CQI, the uplink delay is shown in Fig. 3, Fig. 4 and Fig. 5 respectively.
overall system throughput is lower. The general trend is that the average user throughput decreases
with CQI feedback delay. It is shown from Fig. 3 - Fig. 5
C. Proportional Fair (PF) scheduler that at very low UE speed the average UE throughput is less
A PF scheduler operates in between the Bcqi and RR affected by the CQI feedback delay. This is due to the fact
scheduler, i.e. taken into account the instantaneous channel that UE experience slow varying channel response at very
conditions as much as possible while still satisfy some degree low speed (i.e. 5 Km/Hr) and hence delay has no significant
of fairness among users. In this strategy, the RBs are assigned impact (channel response is relatively same in subsequent TTI)
 1.8GHz, 20MHz, SISO, 100UE/eNodeB, 57 eNo, TS36942 Urban, best cqi
TABLE II 1
AVERAGE T HROUGHPUT OF C ELL E DGE U SERS (M BPS ) FOR RR
ALGORITHM 0.9

Delay (in TTI) 5 Km/Hr 30 Km/Hr 70 Km/Hr 120 Km/Hr 0.8

Average UE Throughput (Mbps)

0 0.02 0.02 0.02 0.02
1 0.01 0.01 0.01 0.00 0.7 UE Speed = 5 Km/Hr
UE Speed = 30 Km/Hr
2 0.01 0.01 0.00 0.00 UE Speed = 70 Km/Hr
0.6
3 0.01 0.01 0.00 0.00 UE Speed = 120 Km/Hr

4 0.01 0.00 0.00 0.00 0.5

0.4
TABLE III
AVERAGE T HROUGHPUT OF C ELL E DGE U SERS (M BPS ) FOR PF
ALGORITHM 0.3

Delay (in TTI) 5 Km/Hr 30 Km/Hr 70 Km/Hr 120 Km/Hr 0.2

0 1 2 3 4
CQI Feedback Delay (in TTI)
0 0.14 0.14 0.14 0.14
1 0.11 0.09 0.03 0.01
Fig. 3. Average UE Throughput versus Feedback delay for best cqi (Bcqi)
2 0.10 0.04 0.01 0.01 scheduler
3 0.09 0.02 0.01 0.00 1.8GHz, 20MHz, SISO, 100UE/eNodeB, 57 eNo, TS36942 Urban, Round Robn
0.26
4 0.08 0.01 0.00 0.00 UE speed = 5 Km/Hr
UE speed = 30 Km/Hr
0.24 UE speed = 70 Km/Hr
UE speed = 120 Km/Hr

on the users average throughput. Very high UE speed (i.e. 0.22

Average UE Throughput (Mbps)

120 Km/Hr) means faster temporal changes in the channel

0.2
and hence even at very low delay average UE throughput is
lowest and afterwards remain unaffected by the delay. For
0.18
average and fast UE speed, it is clear how the average UE
throughput decreases with the increase in delay. Similar trends 0.16
are observed for average cell throughput as shown in Fig. 6,
Fig. 7 and Fig. 8. 0.14

We consider three scheduling algorithms in our simulations:

best cqi, round robin and proportional fair. It is shown in 0.12
0 1 2 3 4
Fig. 3 and Fig. 6 that Bcqi gives superior performance in CQI Feedback Delay (in TTI)

terms of average UE throughput and average cell throughput

Fig. 4. Average UE Throughput versus Feedback delay for round robin (RR)
compare to other scheduling algorithms. This is an expected scheduler
result as Bcqi schedules the users with the best channel 1.8GHz, 20MHz, SISO, 100UE/eNodeB, 57 eNo, TS36942 Urban, Prop Fair
0.55
conditions. Higher throughput of Bqi translates into the fact UE speed = 5 Km/HR
that the distribution of resources among users is not fair. While 0.5 UE speed = 30 Km/Hr
UE speed = 70 Km/Hr
on the other hand, RR and PF average throughput is low 0.45 UE speed = 120 Km/Hr

(compared to Bcqi) but distribution of resources among users

Average UE Throughput (Mbps)

is fair, as in this scheduling algorithm cell-edge users are also 0.4

served (see Fig. 4 and Fig. 5). 0.35

Table II and Table III show the average throughput of cell 0.3
edge users in Mbps for RR and PF respectively. For Bcqi
the average throughput of cell edge users is always zero 0.25

irrespective of mobility. It is clear from Table II and Table III 0.2

that both RR and PF algorithms serve all users irrespective 0.15

of their channel conditions (i.e. CQI values). As a result,
the average UE and cell throughout degrades but fairness 0.1

among users increases. The PF scheduling algorithm tries 0.05

0 1 2 3 4
to balance throughput and fairness and in order to increase CQI Feedback Delay (in TTI)
average throughput it assigns more resources to users with
better channel conditions (i.e. higher CQI). The PF algorithm Fig. 5. Average UE Throughput versus Feedback delay for proportional fair
(PF) scheduler
is intelligently assign resources to users, providing a level
of fairness among users while at the same time providing
minimum quality of service to all users [8].
 1.8GHz, 20MHz, SISO, 100UE/eNodeB, 57 eNodeB, TS36942 Urban, best cqi
90
V. C ONCLUSIONS
80 In this paper, we study the impact of UE speed and CQI
uplink feedback delay on the throughput performance of an
LTE network. By using standard compliance system level
Average Cell Throughput (Mbps)

70

UE speed = 5 Km/Hr simulations, we have shown that in addition to CQI uplink

60
UE speed = 30 Km/Hr
UE speed = 70 Km/Hr
delay the UE speed further degrades system performance. This
UE speed = 120 Km/Hr happens due to the wrong selection of MCS as the CQI value
50 use in the scheduling algorithm may not match the current
channel conditions. We have studied the different scheduling
40 algorithms and calculated average user throughput, average
cell-edge user throughput and average cell throughput under
30 different UE speed and uplink CQI feedback delay. We, finally,
concluded that for an efficient scheduling algorithm at eN-
20
0 1 2 3 4 odeB, UE speed must be taken into account. The development
CQI Feedback Delay (in TTI)
of scheduling algorithm that takes into account the UE speed,
Fig. 6. Average Cell Throughput versus Feedback delay for best cqi (Bcqi)
estimate and correct potential delays in the feedback link is
scheduler the scope of our future work.
1.8GHz, 20MHz, SISO, 100UE/eNodeB, 57 eNo, TS36942 Urban, Round Robn
26
UE speed = 5 Km/Hr
ACKNOWLEDGEMENT
24
UE speed = 30 Km/Hr
UE speed = 70 Km/Hr
The research leading to these results was derived from the
UE speed = 120 Km/Hr University of Greenwich Research & Enterprise Investment
Programme grant under agreement number RAE-ES-01/14.
Average Cell Throughput (Mbps)

22

R EFERENCES
20
[1] T.-T. Tran, Y. Shin, and O.-S. Shin, “Overview of Enabling
Technologies for 3GPP LTE-Advanced,” EURASIP Journal on Wireless
18 Communications and Networking, vol. 2012, no. 1, 2012. [Online].
Available: http://dx.doi.org/10.1186/1687-1499-2012-54
[2] J. C. Ikuno, M. Wrulich, and M. Rupp, “System Level Simulation of LTE
16
Networks,” in Proc. 2010 IEEE 71st Vehicular Technology Conference,
Taipei, Taiwan, may 2010.
14 [3] D. H. Zarrinkoub, Overview of the LTE Physical Layer. John
Wiley & Sons, Ltd, 2014, pp. 13–46. [Online]. Available:
http://dx.doi.org/10.1002/9781118443446.ch2
12
0 1 2 3 4 [4] R. Akl, S. Valentin, G. Wunder, and S. Stanczak, “Compensating for CQI
CQI Feedback Delay (in TTI) Aging by Channel Prediction: The LTE Downlink,” in 2012 IEEE Global
Communications Conference (GLOBECOM), Dec 2012, pp. 4821–4827.
Fig. 7. Average Cell Throughput versus Feedback delay for round robin [5] Q.-T. Quoc-Tuan Vien and H. Nguyen, “CQI Reporting Strategies for
(RR) scheduler Nonregenerative Two-way Relay Networks,” in 2012 IEEE Wireless
1.8GHz, 20MHz, SISO, 100UE/eNodeB, 57 eNo, TS36942 Urban, Prop Fair
Communications and Networking Conference (WCNC), April 2012, pp.
60 974–979.
UE speed = 5 Km/Hr [6] J. Paris, M. Aguayo-Torres, and J. Entrambasaguas, “Non-ideal Adaptive
UE speed = 30 Km/Hr
Modulation: Bounded Signaling Information and Imperfect Adaptation,”
UE speed = 70 Km/Hr
50
UE speed = 120 Km/Hr
in IEEE Global Telecommunications Conference GLOBECOM’04, vol. 6,
Nov 2004, pp. 3828–3834 Vol.6.
[7] L. Zhang, X. Ji, and M. Peng, “A Rank Adaptive Beamforming Scheme
Average Cell Throughput (Mbps)

40 for TDD LTE System with Imperfect CSI,” in 2012 International Confer-
ence on Wireless Communications Signal Processing (WCSP), Oct 2012,
pp. 1–5.
30 [8] S. AlQahtani and M. Alhassany, “Comparing Different LTE Scheduling
Schemes,” in 2013 9th International Wireless Communications and Mo-
bile Computing Conference (IWCMC), July 2013, pp. 264–269.
20 [9] “Evolved Universal Terrestrial Radio Access (E-
UTRA); Radio Frequency (RF) System Scenarios,”
http://www.3gpp.org/DynaReport/36942.htm, accessed: 2014-07-30.
10

0
0 1 2 3 4
CQI Feedback Delay (in TTI)

Fig. 8. Average Cell Throughput versus Feedback delay for proportional

fair (PF) scheduler