Technical Field of the lnvention
The invention relates lo communic€tion netwod(s, and in particular lo a melhod in a
fi6t device for estimating the quality of a signal transmitted from a second device to a
third device.
Backoround to the lnvention
Femtocell base stations in a Long Term Evolution (LTE) communication network
(otheMise known as Home evolved Node Bs - HeNBs - or Enterprise evolved Node
Bs - EeNBs) are small, low-power. indoor cellular base stations for residenlial or
business use. They provide better network coverage and capacity than thal available
in such environments from the overlying macrocellular LT-E netwo* ln addition,
femlocell base stations use a broadband conneclion lo receive data from and send
data back to lhe operator's network (known as "backhaul").
As ferqgrell base stations can make use of the same frequencies as macrocell base
stations in the macrocellular network, and as they are located within lhe coverage area
of one or more fiacrocell base stations in the macrocellular network, it is necessary to
ensure that downlink transmissions from the femtocell base station to mobile devices
(otherwise known as User Equipments - UEs) using the femtocell base station do not
interfere substanlially wilh downlink transmissions from macrocell base stations to
mobile devices using the macrocell base slalions
Typically, thls interference is mitigated by placing a cap on the poller ihal the femtocell
base_station can use to lransmit signals to mobile devices The cap on the power can
be set such that, at a specified pathloss from the femtocell base station (for example 80
dB), a signal received by a mobile device from a macrocell base station would meel a
specified quality level (for example a target srgnal to interference plus noise raiio -
SIR). The determination of the cap is subjecl lo a minimum and maximum power
reslriction on ihe transmission power of the femtocell base slation, for example 0.001
W and 0.1 W respectively
However, lhis approach has limitalons in that the lransmission power of the femtocell
base station will be capped regardless of whelhe. lhere are any mobile devices near lo
the femtocell base station that are communicathg with a macrocell base station and
くυ
り０
10
長）
20
ｒυ
＾Ｚ
30
that ne€d proteciling. This cap cln lead to the data th.oughpul lor motlile devices
co.nmunic€ting with the temtocEll base station b€ing unneceq!g!ry11!!:Led.
ln provlling an approach for s€tting the maximum psmitt€d transmission povrer tor
do$rnlink transmissions ,rom iemtocell base statjons. it is nsc€ssary ,or the ,emtocell
base station to determine it there are nearby mobile devices that need prolecling.
Therefore, there is a need ,oI e metiod in wiich tie femtocell base slation can
determine th€ quality of signals being trans.nitted from a mobile devic€ to anoiher bas€
station.
Summaw of the lnver ion
Therefore, according lo a frst aspect ol the inwntbo, there is provided a method of
estimating a quality o, a signal, the method in a first device comprising measudng a
s€nal transmined fro,n 9 s€cond device lo a tiird device; dele.mining a value o, a
metric i1cm an autocorel€fion function of the me€sured signal; and determinirE an
estimate ol the quality of the sbnal from lhe determined melric
Preferably, the step of m€asuring cornprisgs measuring the signal in the time domain,
ald the step ot determining a value of a metric compris€s determining ihe
autoconelation function ofthe time domain signal and noise.
Preterably, the step of determining a value ot a m€tric comprises d€termining the
autoconelation funclion compfiSeS nomalising ths measured signal to give a sequence
f taking lhe fast Fou.ier lransform of this sequence to give ti determioirE the squared
magnitude of each sample in t; and taking the inveEe fasl Fourier lransform of the
sequence resulting from lhe step of determining lhe squared magnitude to give an
autocorelation sequence a.
ln a prefened embodiment, lhe step of determining a value of a rnetric from an
autoconelation funclion of the measured signal comprises calcllating the magnitude o(
squared magnitude of the autoconelation fundbn.
Preferably, lhe step of detemining a value of a met ic further co.nprises adjusting or
35 zeroing lhe centErl tap in the output of the step of c€lculating
5
10
15
20
25
30
3
ln a furlher embodiment, the step of delermining a value ot a metric furiher comprises
adjusting or zeroing the trap adiacent the central tap in the output of the step of
calcuhting.
ln one embodiment, the Btep o, determining a value of a melric comprises identifying
the tap with the lar96t magnitude of squared magnitude in the taps refilainiog in lhe
output ot the step ol calcuhting: and setting the metric to the value o, said magnitude
or said squarcd magnitude of ihe identified tap.
ln one embodirnent, ths step o, detemining a value o, a metrb further compris€s
adjusting the value of the metric based on the distance ol the irentified tap ho.n lhe
c€ntral tap.
ln anolher embodiment the step of determining a value of a metric further cornprisG
adjusting the value of tha rnstric based on a funclion ol a peak to average porver ralio
of the measured signal.
ln this embodiment, the step ot adiusting the value of the metric based on a tunclion ot
a p€ak to average po$rer ratio ot ihe measured signal preferabty comprises, in the
evenl that the peak to aver€ge power ratio of the rneasud signal is belo$, a threshold
value, adjusting the value ot the metric to a minimum value.
ln one embodimenl, ihe step of dstermining an estimate of the quality o, the signal fro.n
the determined meLic compris€s comparing the dete.mined metric to a look-up table.
ln an altemative embodiment, the step of determining an estimate of the qualily ot the
-- signal from the delermined metric comprises using a cuN+fltting technique to match
the detemined met ic lo a predelermined relationship between value6 for the metric
and the quality of the signal.
Preferably, the step of measuring comprhes measuring a Zadoff-Chu ,eference signal
fansrnined from the second device to lhe third device, and the quality ol the signal is a
signal to noise 6tio.
Preterably, the step ot measuring a Zadoff-Chu referen@ signal composes eslimating
the position ot the Zadofi-Chu reference sgnal in tme
35 ヽ
Preferably, the step of meauring comprises measuring a portion of the Zadofi-Chu
,eference signal.
5 ln one embodimenl, the method further cornprises the step of using a scheduler to
ensure that no signals will be transnitled to the tirst devbe trorn olher devices
associaled lherewilh that might inlerfere with the execution of lhe step of measuring.
According to a s€cond 6psct of the invention. there is provk d a net\,'rort elemenl for
10 uss in a @mmunication net$,ork, the network element bing configured to perrorm the
method describ€d above.
Briet Dqlcriotlon of the Drawlnqs
The invention will now be described in detail, by way of exampl€ onty. with reference lo
15 the follo$ring drawings, in which:
Figure 1 shows an exemplary communication network;
Figure 2 is a lb\x chan ilbstraling a method of setting a maximum permitted
20 t'ansmission pq{er for a femtocell base station;
FEure 3 is a ,low chart illustrating the rnethod of Figure 2 in more detail;
Figures 4(a) and 4(b) are graphs illustrating the autocorehion function tor time
25 domain rererencs signals with lo!, and high signal to noise ratbs respeclively;
Figure 5 rs a graph illustrating a plot of autoconelation function peaks againsl sonal io
noise ratioi
30 Figure 6 is a graph rllustratiog a plot of peak to ave.age porrer ratios against signal lo
noise ratlo;
Figure 7 is a graph illustrating a plot of autoconelation func{ion peaks against signal to
noise ratio in which the scatter has been .educed;
35
10
5
Figure 8 is a flow chart illustrating a method of estimating a signalquality of a relerence
signal in an uplink in ac@rdan@ wilh an exemphry embodiment of the invenlioni
Figure I is a graph illustrating the change in throughput on a macrocell downlink
against femtocell base stalion density in a macrocell sector;
Figure 10 is a graph illustrating the change in throughput on a macrocell downlink
against femtocell base station density for a user equipment at the edge of the
macrocell;
Figure 1l is a graph illustrating the change in throughput on a femtocell downlink
against fernlocell base station density; and
Figure 12 is a graph illustrating the change in throughput on a femtocell downlink
15 against femiocell base station density for a user equipment at the edge of the
temtocell
@odiments
Ahhough the invenlion will be described below with referenc4 to an LTE communication
network and femtocell base stations or HeNBs, it wll be appreciated thal lhe invenlion
is applicable to other types of third or subsequenl generation network in which
femtocell base stations (whether for home or business use), or their equivalents in
ihose networks, can be deployed Moreover, although rn the embodments below lhe
femtocell base stations and macrocell base stations use the same air interface (LTE), it
will be apprecialed thal the invention can be used in a situation in which the macrocell
and femtoc€ll base stations use the same or corresponding frequencies but diflerent air
interface schemes (for example the macrocell base stations could use WCDMA while
the femlocell base stations use LTE)
Figure 1 shows part of an exemplary communication network 2 in which the invention
can be implemented- The communrcation network 2 includes a plurality of macrocell
base stations 4 (only one of which is shown in Figure 1) that each deflne a respective
coverage area - indicated by macrocell 6. ln an LTE communication network, the
macrocell base staiions 4 are refered io as evolved Node Bs (eNBs).
20
25
30
35
5
10
０）
20
25
6
One or more femtocell base stations I (Home eNBs - HeNBs) can be located within the
coverage area 6 of the macrocell base station 4 (although only one temtocell base
slation 8 is sho\rrn in Figure 1), with each femtocellbase station I defining a respeclive
coverage area - indrc€ted by ,emtocell 1 0
It will be appreciated that Figure t has nol been drawn to scale, and that in most realworld
implementations the coverage area 10 of the femtocell base station I will be
signillcantly smaller than the coverage area 6 of the macrocell base stiation 4
A number of mobile devices (UEs) '12 are also located in the communic€tion nelwort 2
within the coverage area 6 ofthe macrocell base stalion 4.
Four mobile devices 12a, 12b, 12c and 12d are each associated with lhe macrocell
base station ,(, meaning that they kansmn and/or receive conlrol signalling and/or data
using the macrocell base station 4. lt will be noled that although the mobile device 12d
is also wiihin the coverage area 10 of the femtocell base station I, it is associaled with
the macrocell base station 4 (this could be due to the signal strength of the macrocell
base station 4 being significantly better for mobile device 12d than the signal skength
of lhe femtoc€ll base siation 8 or the femtocell base slalion 8 could be restricted to
specific subscribeE that don't include mobile device 12d, etc.). Mobile devkles 12a,
12b, 12c and12d are refened to collectively herein as "macro-UEs', as they are the
mobile devices/user equipments (UEs) associated wilh the macrocell base station 4
Two fudher mobile devices, 12e and '12f, are located wilhin the coverage area 10 of lhe
femlocell base station 8 and are curently associated ,^rilh lhe femtoceil base station 8,
meaning that they transmit and/or receive control signalling and/or data using the
femtocell base station 8. Mobile devices 12e and '12f are referred to collectively herein
as 'femto-UEs', as they are the mobile devices/user equipments (UEs) associated with
lhe femtocell base station 8.
As described above, it is necessary to ensure lhat the downlink transmissions from the
femtocell base station 8 to the femto-LJEs 12e and 12f do not prevent nearby macr+
UEs (such as macreuE 12d) frorn being able to successfully receive downlink
transmissions from the macrocell base station 4 A simtlar requkemenl exisls for a
mobile device that is associated with another femlocell base station, rn that the
downlink transmissions from the femtocell base station 8 to the femteuEs 12e and 12f
●υ
００
´０
30
10
15
20
30
7
should not prevent those mobiie devices from successfully receiving the dclwnlink
transmissions from theirfemtoce‖ base stat on
As desc百bed above,thにproЫemヽaddressed h∞ nvenlond ne¨ rks…
cap to the transmission power used by femtoce‖ base slations 8 to transma signals tO
femto UEs This cap is set lo a value that prevents these dcwnlink signals from
causing an undesirable levei orinterrerence to mob‖ e devices that are not associated
with ihe remtOcel base station 8 that are in or near the coverage area 10 of the
femtocel base station 3(such aS mob‖ e device 12d in Figure l) ThiS Cap is applied to
the lransmission pclwer regardless of whetherthere are any moble devices in or nea「
the coverage area 10 oF the femtoce‖ base station 3 (so l wCluld be app‖ ed. for
example,even r mob‖ e device 12d was nol present)
However as iliustrated in Figure 2. it is detennined whether there are any mob‖ e
devices that are not associated with the femtoce‖ base station 8 that require prOtection
from interference caused by dclwnhnk lransm ss ons of the femtoce‖ base slation 8
(Step 101),and the transmission pOwer cap for the femtoce‖ base stamon 8 is set
a∝っrding y(Step 103)
A more detaled method of operaling a femlocel base station 8 is mustrated n Figure 3
ln F19ure 3,steps lll, 113, 117 and l19 cOrrespond to the step of detennining(step
101)n Fture 2
in the fo‖ owing, athough the method w‖ l be described wnh rererence to protecting
mobile device 12d l e a maCrO_UE)that iS assoc ated wnh macrOce‖ base sta“ on 4
from doMlnlink transmissions from the femtoce‖ base station 8, tw]l be app「eciated
that a simila「method can be used to protecl a mob‖ e device that is associated wth
another femtocel base station
in step lll,the remt∝e‖ base sta輛on 8 attempts to identfy r there are any macr● UEs
12 that are receiving downlnk transmissions f「om a macroce‖ base station 4
ln LTE.macr● UEs 12 1ransmn infOnmatiOn to lhe macrOce‖ base statOn 4 beFore,
dunng Or anerthe receipt of a down‖ nk transmission from the macroce‖ base stat on 4,
for example an acknowledgement(ACK/NACK)signal a channel qualw ind cator
(CQり,SOunding signals,data signa s elc TherefOre,the remtα
"‖
base sta“ on 8 can
35
5
25
10
一〇
20
ｒ０
つ４
30
8
monitor uplink channel(s) used by the macro-UEs for these transmissions to determine
if there are any mobile devices nearby that might need protecting from its downlink
lransmissions
ln step l13, l is deternined whether any signais detected in step lll originate from
mobile devices that are not being served by(or aSSOOated wlh)the femt∝ e‖ base
station 8
lf the femtocell base station 8 does not detect any signals from macro-UEs 12, then the
femtocell base station I can assume lhat there are no macreuEs nearby that need
protecting from its downlink transmissions. ln lhis case, in step 115, the maximum
prmitted transmission power for lhe femtocell base station 8 can be set to a high or
relatively high value, for example an upper limil for lhe transmisson power (such as 0.1
W in LTE). The method then returns to step 1 1 1 and repeats periodically.
lf the femtocell base station 8 does detect signals from macro-UEs 12, then lhe method
moves to step 117 in which the remtocell base stalion 8 estimates a quality of a
detected signal. This quahty can be a signal to noise ratio (SNR), a signal to noise plus
interference .atio (SNIR), a signal strength, or any other measure of the qualty of a
tEnsmitted signal. ln some implementations, depending on the way in which the
femlocell base station 8 delects signals in the uplink, the femtocell base slation 8 may
be able to distinguish sgnals from mullrple macro-UEs 12 and can estimate the quality
of each of the signals. Horvever, in altemative implementalions, the femtocell base
station 8 may not be able to distinguish the signals and therefore performs lhe
estimation on the signal with the highest qualityln
a preferred embodiment of the invenlion, the femtocell base station I identifies
characteristics of the Zadoff-Chu reference signal and estimates the signal to noise
ratio (SNR) of this signal This embodiment is described in more detail below with
reference to Figure 4 lt will be noled that in this embodimenl lhe femtocell base
station I does not distrnguish between signals from multiple macro-UEs 12 and
lherefo.e estrmates the SNR for the signalwith the highesl quality
ln an alternative implementation, the femtocell base station 8 detects and decodes the
35 data in the uplink and determines a quality of the data signals. It will be appreciated by
´０
10
lhose skilled in the art that altemative techniques can be used by the femtocell base
stalion I io delermine a quality ofthe signals in the uplink.
The femtocell base station 8 then compares the estimated quality (or the highest
estimated quality rf the femtocell base slation 8 can estimate the quality for multiple
signals) with a threshold value (step '119) ln a preferred implemenlaUon where the
quality is a signal lo noise ratio, the threshold can be a value rn the range of 10 dB to
30 dB.
It will be noted lhat a macro-UE 12 will need most proteclion from the downlink
transmissions of the femtocell base stalion I when it is near lo lhe edge of the
coverage area 6 of the macrocell base station 4, as lhe downlink signals received al
the macro-UE 12 from lhe macrocell base station 4 will be relatively weak. ln this
s ualion, lhe macro-UE 'i2 will need to be transmitting its uplink signals al a relatively
high power (due to its distance from the macrocell base station 4). By estimating a
quality of the uplink signal (which will be affected by the transmission power of the
macro-UE 12d and its proximity to the femtocell base station 8), the femtocell base
station 8 can determine whether, and/or the extent to which, the macrGUE 12d needs
protecting from the downlink transmissions of the femtocell base station L
Therefore, if the estimaled quality exceeds the threshold value then the femtocell base
station 8 assumes that the macro-UE 12d that originaled the signal needs significant
protec{ion from the downlink transmissions of the femtocell base station 8, and the
maximum permitled lransmission povre. for lhe femtocell base station I should be set
at a low or relatively low value (step 121) For example, the maximum permitled
transmission power can be set to a lower limit for the lransmission power (such as
0 001 W in LTE).
ln one rmplemenlation, the femtocell base station 8 sets the maximum permitted
kansmission power such lhat, at a specified pathloss from the femtocell base station 8
(tor example 80 dB), a signal received by lhe macro-UE 12d from the macrocell base
stalion 4 meets or is eslimated to meet a specified quality level (for example a target
signal to interference plus noise ralio - SINR), as in a conventional network
The method then returns to step 1'11 and repeats periodrally.
´０
20
25
30
35
10
くυ
20
25
30
10
lf the estimated quality does not exceed the threshold value lhen the femtocell base
station I sets the maximum permitted transmission power to an intermediate value that
lies beh,veen an upper and lower limit for the transmission po er (step 123). Thus, ihe
femtocell base stalion I provides some prolection for the macreuE 12d, while allowing
downlink transmissions from the femtocell base station I to be transmitted at a higher
power than conventional techniques permit In this way, the data throughput for femto-
UEs 12e and 12f can be improved over the conventional technique
ln a preferred implemenlation, the iniermediate value for the maximum pemitted
transmission power is selected based on the d ference between the estimated quality
of the signal and the threshold value. ln partlcular, the value for the maj(imum
permitted transmission power can increase in proportion to the difference between the
estimated quality of the signal and the threshold value (up to an upper limit, if
applicable). ln a preferred embodiment where the quality is a signal to noise ratio, f
the estimated SNR is 5 dB below the threshold value, then the maximum permitted
lEnsmit power c€n be set to be 5 dB above lhe lod or relatively low value, subject to
the upper limit on the maximum permitted transmit power
Again, the method returns to step 111 and repeats periodically
ln one implementation of the invention, steps 113 and 117 can be combined, in that lhe
lemtoc€ll base station 8 estmates a quality (such as the SNR) of a signal in the uplink
and if the estimated quality is above a particular threshold, then a deteclion of a macro-
UE 12 is assumed to have been made. This threshold could be the same or differenl
to the threshold used in step 119.
It will be appreciated that a macreuE 12d may move into the vicinity of the femlocell
base station 8 (i.e into or near to the coverage area 10 of the lemtocell base station 8)
without needing to transmit anything to its associated macrocell base station 4 (for
example if lhe macro-UE 12d is noi receivng any downlink lransmissions from the
macrocell base station 4), which means that the Iemtocell base station 8 will not be
able to detecl lhe macro-UE 'l2d in slep 111
However, as the macro-UE 12d may need to monitor downlink conlrol channels from
the macrocell base station 4 (for example a broadcast channel - BCH, or a physical
downlink control channel - PDCCH), i is necessary lo make sure that the macro-UE
35
10
一〇
20
25
11
12d is able to receive lhese downlink transmissions. Allhough these channels are
designed to be rehtively robust againsl interference, the femtocell base station I may
still interfere with these channels if the transmission power is sufficiently high.
Therefore, in one implementation, the femtocell base station I periodically or
inlermittently sets the maximum permitted transmission power to the lower limit, in
order to provide lhe maximum prolection for any macro-UEs 12d in its vicinity,
inespective of whether the femtocell base station 8 detecis any signals in steps 11'l
and 113 For example, the femtocell base station 8 can set the maximum permitted
transmission power to the lower limil for i00 milliseconds eve.y'l second This will
provide opportunities for any macro-UEs 12d that are not transmitting any uplink
signals lo listen for dolvnlink transmissions from the macrocell base station 4.
ln an alternative implementalion, the femtocell base station 8 c6n set the maximum
permitted transmission power to the lower limit wheneverthe temlocell base station 8 is
transmiffng signals at the same time that the macrocell base station 4 is transmitting
control channel signals. ln partrcular, the femlocell base station I will typicalty be
synchronised with the macrocell base station 4 and the control channel signals will be
sent at predetermined times and on predetermined resource blocks (RBs), so ihe
femtocell base station will know when the macrocell base station 4 will be transmitting
the control channel signals- For example, in LTE, some control channel signals are
hansmitted once everlms (e.9. PFICH, PDCCH), wilh the first four of fourleen
symbols transmitled per 1ms canying control channel signals. Olher conlrol channels
(e g PBCH, PSCH) are sent less frequently and use approximately seven symbols out
of every 140 symbols and a subset of the available resource blocks.
Eslimalion of the oualitv of an uolink reference sional
As described above, in a preferred embodiment of the invention, the femtocell base
station 8 identilies characteristics of the Zadoff-Chu reference sgnal and estimates the
signallo noise ratio (SNR) ofthis signal.
Unlike WCDMA networks, in LTE the characteristics of uplink relerence signals are
significantly different to the characteristics of both data transmissions and thermal
noise. This method explorts dfferences in the autoconelation function between a
podion ofthe time domain reference signal and (liltered) Gaussian noise
30
一〇
10
●０
20
25
30
12
For an uplink reference signaloccupying a small number of frequency domain rcsource
blocks, it would be expected that the autoconelation function with high SNR would
deviate from thal due to (filtered) Gaussian noise However, even with a wideband
speclrally flat reference signal, such as 50 resource blocks (lhe maximum for a 1O MHz
system), the autocoraelation Iunclion of a portion of the time domain reference signal
deviates from the fihered Gaussian noise case.
This is true for all the Zadoff-Chu basis sequences, although the nature of the
autoconelation funclion does depend on the particular Zadoff_Chu basis sequence An
example of the autoconelation ,unction for low and high SNR cases with 50 resource
blocks is shown in Figures 4(a) and 4(b) respectively.
It can be seen in Figure 4 that lhe low SNR case is dominated by the autoconelation
function of the fiftered Gaussian noise, while lhe high SNR case is dominaled by the
adoconelation function ot the referenc€ signal
Figure 5 shows the resulls of a simulation in which the autocorrelation peaks from a
shgle reference signal, excluding the centraltap, is plotted against the SNR. This ptot
$.as obtained over a range of different reference sighal paramelers, numbers of
resource blocks, numbeE of macro-UEs, SNRS from each macro_UE and ftequency
resource assignments_ The simulation atso included fading effects.
Thus, il c€n be seen from Figure 5 that this metric, based on the autocorrelalDn
function, c€n be used to estimate or predict the SNR in many cases Ho\/ever, lhere
are a number of poinls in the plot where although lhe SNR is high, the metric remains
low. This scatter to the right hand side of the plot is potentially problematic, since in
these cases nearby macro-UEs might not be protecled by the femtocell base station I
This scatler can be due to fading as well as diffeaences between the autocorrelalion
functions of the different Zadoff-Chu basis sequences
An altemative class of metric for the estimation of the SNR c€n be based on the
statistics of the time domain waveform. One simple metric is the peak lo average
power ratio (PAPR). H'gh SNR reference signals should have low pApR, whereas
Gaussian noise has a relativety high pApR
‘）
００
10
15
20
〓０
＾Ｚ
30
13
Results Ior this metric (in linear units) are shown in Figure 6 and it can be seen ihat
there is an even rarger scatter apparent in the pApR metric than the autoconeration
metric, and as such lhe pApR metric (and other metrics based on statistics of the
power) are less aflractive for estimatmg the SNR of the uplink reference signal.
However, il has been observed that the scattering between the autocorelation and
PAPR metrics ls independent, i.e. for the problematic points with high SNR but
abnormally low autocorrelation metric, the pApR tehds to remain low (as expected for
high SNR signals). For such points, the aulocorrelation metric c€n be adjusted
(upwards). This approach can be used to reduce the scattq in the autocorrelation
metric, and therefore improve the esttmation of the SNR. For example, the pApR p
(in linear units) is tess than 3, then a minimum value can be applied to the metric, thas
minimum value being given by 4OO+(3-p).50.
Two additional approaches for fu(her reducing the scatter in the autocorelation metic
have been identified.
Firstly, as the autocorrelation peaks of lhe reference signals tend lo reduce rn
magnitude with distance from the marn central peak, then some shaping of the
auloconelalion function can be applied To avoid an increase in the.false deteclton"
rate, il is impoftant that lhis is only done for samples in the autocorrelation function
which are already significanfly above the noise tevet _ and so a threshotd is applied
prior to applying this shaping For exampte, if the metric is greater than 120 and the
offset from lhe centre tap is n then the metric can be increased by 0.6r.
Secondly, the scatler can be reduced by obtaining resufts over muhiple measurements,
for example by laking the maximum mekic oblained from a set of four or eighl
measurements
By using all of these lechniques, the scatter in the autocorrelation metric is significanily
reduced Figure 7 iltustrates the resulting relationship between the autocorelaton
melric and the SNR
The femtocell base station 8
auloconelation function and lhe
can make use of the relationship between the
SNR to determine the SNR of an uptink signal. A
〓０
＾０
10
一〇
20
〓υ
＾Ｚ
30
14
method of estimating the SNR of the Zadoff-Chu reference signal in accordance with
an embodiment of the rnvention is shown in more detait in Figure 8
Firstly, the femlocell base station I obtains a ,rough, synchronization to the macrocell
(via a network monitor mode, or, if the standards altow, via macrocell timing
measuremenl reports included in mobile device measurements, or via the X2
interface).
This rough synchronization alows the femlocell base station 8 to estimate roughty
where in time the uptink reference signals from rnacreuEs are likely lo be. ln nearly all
cases, this is the centre symbol in the O.Sms uptink suEframe.
It will be appreciated that this estimation wilt be subjecl to some error due lo
propagation delay from the macrocell base station 4 and lhe timihg advance used by
macro-UEs 12 ln the case of oveFthe-air synchronization, which ls assumed
hereafler, lhe error will be up to one macrocell round{rip propagation delay, which for a
cell of skm is 33us which is roughty half the duraton of an orthogonal frequency
division multiplexing (OFDM) symbol. The enor means that signals received from
macreuEs 12 may arrive earlier than expected al the femtocell base station 8.
Therefore, in step 201 of Figure 8, the femlocell base station 8 measures or captures a
portion of the uplink reference symbot to give a time domain reference signal. For
example, the femtocell base station I obtains the time domain reference signal from
the firs| 512 samples of lhe reference symbol (assuming a .lO MHz bandr,vidth with
1024 samples plus a cyclic prefix per OFDM symbol) Despite the timing uncertainty
for over-lhe-air synchronization, this captured ponion of the reference symbol Should
only contain reference signal samples kom macro-uEs .12 lhat are near lo the
femlocell base sialron 8 (r e. there shoutdn't be any s€mples of data symbols).
ln this step, a scheduler in the femtocelt base stalion g may be used to ensure thal
there will be no uplink transmissions from femteuEs 12 to the femtocell base station 8
lhat might interfere with this measuremeni
ln step 203, the femtoce base station g determines the autoconelation function for the
35 time domain reference signal and (fittered) Gaussian noise.
10
´０
20
25
30
t5
ln one implementation, the femtocell base station 8 does this by normalizing the
captured time domain signal to give unit power, with the resuhing sequenc€ being
denoted r, taking the fast Fourier transform (FFT)ofthis sequence to give f, calculating
the squared magnitude (12+02) for each sample ol f and taking the inverse FFT of lhe
resulting sequenc€ to give the aulocorrelation sequence a.
As the autoconelation sequenc€ a delermined in step 203 is symmetrical (see Figure
4), only half of the samples in a need to be retained by the femlocell base stalion 8 for
further processing.
ln step 205, the femtocell base stalion I lakes lhe magnilude (or, in allernalive
implementations, lhe squared magnitude) of sequence a and lhen, in step 207, adjusts
or zeros the central lap (conesponding lo zero lime lag in lhe autoconelation funclion)
ll may also be necessary to adjust o. zero the tap adjacent to the cenkal tap if this tap
is significantly influenced by filtering in the receive path. Such Iiltering has a fixed
characteristic so the decision as to adjust or zero this tap is a design decision.
Then, in siep 209, the femlocell base slation I finds lhe tap with the largest magnitude
(or squared magnitude) in the remaining taps, and s€ts the value of a metric m to this
magnitude (or squared magnitude).
The femtoc€ll base station 8 can lhen determine the signal to noise ralio of the uplink
reference signal using this metric (step 21 1) The value of lhe SNR for the determined
metric m c€n be determined from the relationship shown in Figure 5 or Figure 7, for
example using a curve-f('ting technique or a look-up table.
As described above, the accuracy of the SNR estimation can be improved by
considering the PAPR of the signal, shaping the autocorrelation function based on the
distance of the peak used to determine the melric from the central tap and/or the metric
may be estimaled from signals received in muftiple time slots
Therefore, the metric m may be adjusted as a funciion of distance from the central lap
for example by applying a simple linear funclion to the metric m determined in step
35 209. This linear function can be as described above.
10
一〇
20
25
30
16
Additionally or altematively, the metric m may be adjusted as a function of the peak to
average power ralio of the captured portion of the uplink reference symbol. Specifically
if the PAPR is below a threshold (for example 3 in linear units) then a minimum vatue
can be imposed on the melric (again this can be a simple lingar function of pApR).
Again. this linear function can be as described above.
Again, additionally or attematively, the metric m or SNR may be eslimated trom uplink
reference signals captured in multiple time slots and, for example, the highest value of
the SNR obtained from these measurements can be used by the femtocell base station
I to adjust its maximum permitted transmission power.
Figures I to'12 illustrate the performance benefrls of the approach described above.
Figure I illustrates hor., the data throughput on a downlink fro.n a nEcrocelt base
station is affected by an increasing number of aclive femtocell base stations within the
coverage area of the macrocell base slation for both a conventional lixed power cap
and the scheme described above. ln particular, it can be seen lhat there is a negligible
differenc€ in the data throughput between the conventionat scheme and the scheme
described above.
Figure 10 illustrates how the data throughput on a downlink from a macrocell base
station to cell edge (5 percentile) macro-UEs is affected by an increasing number of
active femtocell base stations within the coverage area of the macrocelt base station
for a convenlional scheme and a scheme as described above. Again, there is almost a
negligible difference between the two schemes.
Figure 1'l plots the data throughput on a downlink from a femtocell base station againsl
the number of active femtocell base slalions within the coverage area of the macrocelt
base staiion for both a conventional fixed power c€p and the scheme according to the
invention ll can be seen thal lhe scheme described above provides an approximale
increase in data lhroughput of 5 Mb/s regardless of the number of active femtocell base
stations, which is roughly equivalent to an improvement of 25% rn the dala throughpui
Figure '!2 plots the dala throughput on a downlink from a femtocell base stiation to ce
edge (5 percentile) femtcuEs against the number of active femtocell base stations
within the coverage area ofthe macrocell base siation for a conventional scheme and a
35
10
´５
20
17
scheme as described above. lt can be seen lhat for cell edge (5 percentile) femto-UEs
the scheme described above provides an approximale increase in data throughput of
190 kb/s regardless of lhe number of active femlocell base stations, which translates to
an eight-fold increase in the data throughput.
Therefore, these graphs indicate that the adaptation of the maximum permitted
transmission power according to the invention provides perfo.mance benefits ror femto-
UEs over the conventional fixed maximum pemitted transmission power scheme, while
offering the same protection to the macrocell base stalion downlink.
Ahhough lhe invention has been described in terms of a method of estimating a signal
quality, il will be appreciated that the invention can be embodied in a femtocell base
statjon lhal comprises a processor and transceiver circuitry configured to perforn the
described method
Furthermore, while the invention has been presented as a method in a femlocell base
station of estimating a quality of a signal kansmitted from a macrGUE to a macrocell
base station (or from a femteuE to another temtocell base station) that allows the
femtocell base station to control its maximum permitted lransmission power, it will be
appreciated that the signal quality estimated using the method according to lhe
invention can be used ,or other purposes, and c€n be perfomed by elemenls in a
communication network other than femtocell base stations, such as macrocell base
stations (eNBs) or mobile devices-
While the invenlion has been illustrated and described in detail in the drawings and
foregoing description, such illustrahon and descriptlon are lo be considered illustrative
or exemplary and not restriclive; lhe invenlion is not limited to the disclosed
embodiments.
Variations to the disclosed embodiments can be understood and effecled by those
skilled in the art in practicing the claimed invention, from a study of the drawings, the
disclosure, and the appended claims. ln the claims, the word 'comprising" does not
exclude other elements or steps, and the indefnite article "a" or'an" does not exclude
a plurality. A single processor or other unit may fulfil the functions oI several items
35 recited in the claims. The mere fact lhat certain measures are recited in mutually
different dependenl claims does nol indicate lhat a combtnation of these measures
30
25
18
cannot be used to advantage. A computer program may be stored/distributed on a
suitable medium, such as an optical slorage medium or a solil-state medium supplied
together with or as pad of other hardware, but may atso be distributed in olher forms,
such as via the lnternet or other wired or wireless telecommunication systems. Any
reference signs in the claims should not be construed as limiling the scope.

We Claim:
T An apparatus for a wireless communications nelwork compristng al least one
processor to
detenlrine tl there are any transrntssions from a user equipment UE to a ftrst base
slahon by monrlonng Lrpltnk reference stgnals between the UE and the frrst base stalton
measure a quality of any detected uplink reference signals belween the UE and lhe
first base slalion to delermine if the UE is nearby the second base station, where,n the
qua(y measurement comprises deterrntntng a value of an autocorreiatton funclron ol the
detected upllnk reference srgnal ln a trme domain and
conlrol rn lhe second base slalion tnlerference caused lo communicaltons belween
the frrst base slation and the UE by communications of the second base stalion to mitigate
interference wilh lhe UE. the controt dependrng upon the measured quality of the detected
uphnk reference signals between lhe UE and the first base station
2 Apparalus as clarmed rn clarm I wheretn the at least one processor is to lnvoke
down[nk power controL tn the second base station in response to delecting the presence of
the UE needrng prolectron based upon ihe measured quatrty of the detected up link reference
3 Apparatus as carmed in caim 1 wherein determinal on of the value of
autocorrelalron functron by lhe at least one processor compnses taking a Fourier Transform
of a sequence based on the measured uplink reference signal
4 Apparatus as clarmed in clarm 1 wherein the at leasl one processor ts to eslimale
the qualily of the detected upllnk reference stgnal by calculaling a magnitude or squared
magnitude to determtne lhe value of autocorrelation function
5 Apparalus as clarmed in clarm 1, wherein the uptink reference signal is a Zadhoff-
Ch! reference signa
6 Apparalus as clarmed tn any one of lhe precedtng clatms wheretn the ftrst base
slatron rs a macro evo ved Node B, macro eNB and the second base statio| is a Home
ENB
7 Apparatus as ctaimed in clarm 6, wherein lhe Home eNB is reslricted to specific
subscnbers excluding the UE
-20-
I Apparatus as claimed in claim 1 or claim 6' wherein the apparatus is implemented in
processing circuitry in the second base station
I A method for controlling interference a wireless communications network comprising
a first bases statlon' a second base stalion and a user equipmenl' UE' the method
compilslngl
monitoling uplink relerence signals beNveen the UE and lhe firsl base station lo
delermine if communications between the UE and the flrst base statlon need protedon from
lransmsslons of the se@nd baEe station'
determinlng a metric of any uplink reference signals detected' the metric compnslng
hking a value of an autocorrelation coresponding to the detected uplink reference signals;
and
controlling interference to communrcations between the first base station and the UE
by communcations of the seconcl base statlon to mitigate interference with the UE the
controlbeingperformedinthesecondbasesiationdependingonlhedeterminedmetricof
the detected uplink reference signals
)s invoking
10 The method of claim 9' wherein the controlling interference comprs(
downlink power control in the s€cond base station in response to detecling the presenc€ of
the UE needing protection based upon the measured quality of the detected uplink reference
signals.