Abstract;

Introduction;

The radio resource allocation algorithms:

◦ Fair Allocation scheme;

◦ Proportional Allocation scheme.

Simulator description;

Simulation results;

Conclusion;

References.

In this paper is proposed an algorithm which distributes

the available bandwidth proportionally to the channel

conditions while meeting the SINR target by means of

power control.

The results show that the proposed algorithm, when

combined with slow power control, is able to improve

the average cell throughput by approximately 23% with

almost no reduction of the 5% outage average user

throughput.

In the E-UTRAN system the users of the same cell are

orthogonal with each other thus the only interference

present is from other cells.

In this paper, which is intended to be an initial

evaluation of the system performance of E-UTRAN,

the MCS is assumed to be fixed, thus the focus is on

the performance of ATB.

This algorithm equally distributes the available PRBs among the users in the cell. The number of allocated PRBs per UE changes only when the number of UEs in the cell changes because of a handover. The closed loop PC adjusts the transmit power of the UE, in a fast or slow basis, depending on the received SINR in order to match the SINR target. If the received SINR at the Base Station (BS) is less than the SINR target, a power-up command of 1 dB is given to the UE while if the received SINR is greater than the SINR target, a power-down command of 1 dB is given.

This algorithm dynamically adapts the user’s bandwidth to the changing channel condition while trying to match the SINR target by means of either SPC or FPC.

After an initialization phase, in which each UE receives an equal number of PRBs,

the SINR is measured and its ratio to the SINR target is calculated. Using such a

ratio between measured SINR and SINR target, and the number of previously

allocated PRBs, the number of PRBs necessary to match the SINR target while

transmitting at lowest power is calculated.

Similarly, the number of PRBs

necessary to match the SINR

target while transmitting at

highest power is calculated.

These minimum and maximum

numbers of requested PRBs

carry information on the current

channel conditions the UE is

experiencing within the

allocated PRBs.

To begin with, the packet scheduler tries to guarantee the minimum requested

allocation of PRBs to each user. In case the number of available PRBs is less than

the requested number, the allocation will take place in proportion to the minimum

request from each user. If the number of available PRBs is higher than the

minimum requested, the extra PRBs will be distributed among the users

proportionally to the highest request from each of them.

Finally, once the allocation is performed, the power is scaled according to the ratio

between current and previous bandwidth in order to keep power spectral density

constant. Also in this scheme the closed loop PC adjusts the transmit power of the

UE in order to match the SINR target. The aim of this radio allocation scheme is to

allocate more PRBs to users in the BS vicinity which have a lower power spectral

density, than the ones at the cell boundary which have a higher power spectral

density.

This results in a reduction of intercell

interference and consequently in increased

system capacity. The underlying assumption

is that the channel, on a steady-state, will

have a similar behavior over consecutive

PRBs.

ELIISE - Efficient Layer II Simulator

for E-UTRAN, is a multi-cell, multi-

user, dynamic system level simulator to

study advanced RRM in uplink. The

functionalities include channel model,

mobility, handover, Automatic Repeat-

reQuest (ARQ), PC and ATB.

The simulated network layout assumes

a hexagonal grid with 8 BSs and 3

sectors per BS with a corner-excited

structure.

This figure shows the distribution of the uplink SINR per TTI for all

the UEs in the system. The FA algorithm shows a better matching of

the SINR target than the PA algorithm because the allocation of the

same PRBs to the same UEs over subsequent TTIs makes it easier to

track the fading channel and the more stable intercell interference.

As expected FPC offers

better matching properties

than SPC for both schemes.

pdf of the instantaneous SINR for all UEs measured on a TTI basis

This figure shows that all

UEs have higher time

average SINR when using

PA as compared to FA for

both SPC and FPC. In the

PA case the UEs at the

cell edge receive less

PRBs thus it is easier for

them to reach the SINR

target.

CDF of the time average SINR per UE

The distribution of the

average UE throughput is

shown in this figure. Here

it can be seen how the PA

algorithm trades a small

loss in the lower end of

the throughput range for a

significant improvement

in the higher range.

CDF of the time average throughput per UE

The distribution of the

average number of PRBs

per UE, given in this

figure, shows that the PA

mechanism is a channel

aware scheme, thus it is

able to adapt to the

channel conditions as

shown by the greater

variability in the number

of allocated PRBs. CDF of the time average PRBs per UE

The cell throughput results are shown in this figure. The cell throughput of the

PA scheme is higher than the FA scheme by 23% for the case of slow PC and

15% for the case of fast PC. At 5% outage of the average user throughput the

PA scheme shows a loss of approximately 2% in case of slow PC and a gain of

approximately 18% in case of fast PC. Thus the fairness for the users in the

lowest throughput range is preserved.

A possible improvement for the

Quality of Service (QoS) of those

users could be obtained by setting a

minimum requirement on the bit rate.

It is interesting to notice that the gain

from FPC decreases from 22% to

15% when moving from FA to PA.

Average cell throughput and 5% outage average user throughput

On the left, it is clear

that FPC is not able

anymore to guarantee

a SINR matching

better than SPC. On

the right, the gain of

FPC is reduced to

7% for the case of

FA.

pdf of the instantaneous SINR for all UEs (measured on a TTI basis)

and average cell throughput for the case of randomized PRBs

allocation

In this paper have been presented studied the performance of two resource allocation algorithms in combination with SPC and FPC for UTRAN LTE, assuming fixed MCS and SINR target. The results obtained show that the PA algorithm increases the average cell throughput by 23% for the SPC case, while slightly reducing the throughput of the UE in disadvantaged channel conditions.

It is shown to depend on the allocation scheme and it decreases from 22% to 15% for PA because of higher interference variations.

Future studies will concern the integration and performance evaluation of other LA functionalities.

[1] 3GPP TR 25.814 V7.0.0 (2006-06), “Physical Layer Aspects for Evolved UTRA”.

[2] A. Toskala and P.E. Mogensen, “UTRAN long term evolution in 3GPP”, International Symposium on Wireless Personal Multimedia Communications, 2005.

[3] A. Toskala, H. Holma, K. Pajukoski and E. Tiirola, “UTRAN long term evolution in 3GPP”, IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, Sep. 2006.

[4] A. Pokhariyal, T.E. Kolding and P.E. Mogensen, “Performance of Downlink Frequency Domain Packet Scheduling for the UTRAN Long Term Evolution”, IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, Sep. 2006

[5] J. Lim, H.G. Myung, K. Oh and D.J. Goodman, “Channel-Dependent Scheduling of Uplink Single Carrier FDMA Systems”. IEEE Vehicular Technology Conference, Sep. 2006

[6] H. Holma and A. Toskala, Eds., “WCDMA for UMTS”, 3rd ed., John Wiley & Sons.

[7] T. Hyt¨onen, “Optimal wrap-around network simulation”, Helsinki University of Technology Report, A432, 2001.

[8] 3GPP TR 25.943, “Deployment aspects”.