Femtocell Solution Radio Access Network Features

NEC Femtocell Solution
Radio Access Network Features

Ref: NEE0858554
Date: 5
th
Jan. 2011

NEC Confidential 5
th
J an. 2011
Copyright 2011 by NEC Europe. All Rights Reserved.
1

Copyright
This material is protected by the copyright laws of the United Kingdom and other countries. It may not be
reproduced, distributed, or altered in any fashion by any entity, except in accordance with applicable
agreements, contracts or licensing, without the express written consent of NEC Europe.
Revision History
Date Issue Author Amendment Details
21/08/2008 1.00 NEC First Release
16/07/2009 2.00 NEC Update

NEC Confidential 5
th
J an. 2011
2

TABLE OF CONTENTS
1 GLOSSARY ............................................................................................................... 3
2 3GPP PROCEDURE AND RADIO NETWORK FEATURES .................................... 7
2.1 Mobility Management ..................................................................................................................... 7
2.1.1 Location Area and Cell ID assignment for Mobility Management .......................................... 7
2.1.1.1 Concept of Super Cell and Super LAC ............................................................................... 7
2.1.2 Cell Reselection .................................................................................................................... 13
2.1.2.6 Configuration for Cell Selection/Reselection ..................................................................... 16
2.1.3 Handover............................................................................................................................... 19
2.2 Radio Resource Management ..................................................................................................... 22
2.2.1 Interference Management ..................................................................................................... 22
2.2.2 Power Control ....................................................................................................................... 23
2.2.3 Admission Control ................................................................................................................. 29
2.2.4 Capacity and Multi-RAB Combination .................................................................................. 32
2.3 HSDPA ......................................................................................................................................... 33
2.3.1 Expected Throughput ............................................................................................................ 33
2.3.2 HSDPA Scheduler in FAP..................................................................................................... 34
2.4 HSUPA ......................................................................................................................................... 34
2.4.3 HSUPA New Channel ........................................................................................................... 35
2.4.3 HSUPA Uplink Scheduling .................................................................................................... 35
2.4.4 HSUPA HARQ ...................................................................................................................... 36
2.4.5 HSUPA Power Management ................................................................................................ 36

NEC Confidential 5
th
J an. 2011
3

1 Glossary
Definition Abbreviation
Second Generation 2G
Third Generation 3G
Third Generation Partnership Project 3GPP
16 Quadrature Amplitude Modulation 16 QAM
Absolute Radio Frequency Channel Number ARFCN
Access Overload Control ACCOLC
Adaptive Multi-Rate AMR
Application Programming Interface API
Broadcast Channel BCH
Block Error Rate BLER
Base Station Identity Code BSIC
Carrier to Interference Ratio C/I
Cell Global Identity CGI
Channel Quality Indication CQI
Code Division Multiple Access CDMA
Common Pilot Channel CPICH
Dedicated Channel DCH
Downlink DL
Dedicated Physical Control Channel DPCCH
Dedicated Physical Data Channel DPDCH
Digital Subscriber Line DSL
Bit Error Rates E
b
/N
o

E-DCH Absolute Grant Channel E-AGCH
Enhanced Dedicated Channel E-DCH
Enhanced Dedicated Channel-Dedicated Physical Control Channel E-DPCCH
Enhanced Dedicated Channel-Dedicated Physical Data Channel E-DPDCH
Enhanced HARQ Acknowledgement Indicator Channel E-HICH

NEC Confidential 5
th
J an. 2011
4

E-DCH Relative Grant Channel E-RGCH
Femto Access Point FAP
Fractional Dedicated Physical Channel F-DPCH
Field Programmable Gate Array FPGA
Generic Access Circuit-Switched Resource GA-CSR
Generic Access Network Controller GANC
GSM EDGE Radio Access Network GERAN
Global Hand-in Channel GHC
Hybrid Automatic Repeat Request HARQ
High Speed Downlink Packet Access HSDPA
High Speed Downlink Shared Channel HS-DSCH
High-Speed Uplink Packet Access HSUPA
International Mobile Subscriber Identity IMSI
IP Network Controller INC
Quality of Service QoS
Quadrature Phase-Shift Keying QPSK
Location Area LA
Location Area Code LAC
Location Area Identifier LAI
Location Area Update LAU
Media Access Control MAC
Mobile Switch Centre MSC
Non Access Stratum NAS
Node B Application Protocol NBAP
Orthogonal Variable Spreading Factor OVSF
Power Control PC
Packet Data Protocol PDP
Protocol Data Unit PDU
Primary Common Control Physical Channel PCCPCH

NEC Confidential 5
th
J an. 2011
5

Primary-Common Pilot Primary Common Pilot Channel P-CPICH
Public Land Mobile Network PLMN
Packet Temporary Mobile Subscriber Identity P-TMSI
Radio Network Controller RNC
Radio Radio Control RRC
Primary Synchronization Channel P-SCH
Routing Area RA
Radio Access Bearer RAB
Routing Area Code RAC
Routing Area Identifier RAI
Radio Frequency RF
Radio Network Controller RNC
Random Access Channel RACH
Radio Access Technology RAT
Routing Area Update RAU
Radio Frequency RF
Radio Resource Connection RRC
Radio Resource Management RRM
Received Signal Code Power RSCP
Subchannel Number SCN
Serving GPRS Support Node SGSN
Shared Hand-in Channel SHC
System Information Block SIB
Signal to Interference Ratio SIR
Subscriber Identity Module SIM
Super Location Area SLA
Small Medium Enterprises SME
Serving Radio Network Subsystem SRNS
Secondary Synchronisation Channel S-SCH

NEC Confidential 5
th
J an. 2011
6

Temporary Mobile Subscriber Identity TMSI
Transmit Power Control TPC
Transmission Time Intervals TTI
Transmit Tx
UTRA Absolute Radio Frequency Channel Number UARFCN
Universal Equipments UEs
Uplink UL
Universal Mobile Telecommunications System UMTS
UMA Network Controller UNC
UMTS Terrestrial Radio Access Network UTRAN
UMTS Terrestrial Radio Access Network UTRAN
Visitor Location Register VLR

NEC Confidential 5
th
J an. 2011
7

2 3GPP PROCEDURE AND RADIO NETWORK FEATURES
2.1 Mobility Management
2.1.1 Location Area and Cell ID assignment for Mobility Management
To facilitate mobility management functions in the UMTS Terrestrial Radio Access Network (UTRAN), the
coverage area of an Mobile Switch Centre (MSC)/Visitor Location Register (VLR) is split into logical
registration areas called Location Areas. Similarly, the coverage area of a Serving GPRS Support Node
(SGSN) is split into logical registration areas called Routing Areas. Universal Equipments (UEs) are
required to perform a location update with the VLR each time the serving location area changes and to
perform a routing area update with the SGSN each time the serving routing area changes.
In the Femto Access Point (FAP) network, one of the key challenges is detection of UE by the FAP. The
UE can camp on a FAP via its cell selection procedure. However, if the UE is in idle mode, there will be
no messages exchanged between the UE and the FAP, thus making it difficult for the FAP to detect the
presence of the UE. In order to trigger an initial message from UE upon its camping on a specific FAP,
the FAP will need to be assigned a different Location Area (LA) from that of the neighbouring macro cells.
This will result in the UEs Mobility Management (MM) layer triggering a Location Update message to the
core network via the camped cell i.e. FAP. The FAP also gets chance to inform the UE at earliest time,
through Location Update accept message or local access control, if the UE is an authorised user.
The Location Area Code (LAC) space is limited to a theoretical maximum of 64K values (due to the
limitation of a 16 bit LAC attribute). As a result, the LAC allocation must provide a mechanism for the re-
use of LAC values for multiple FAPs to scale beyond the 64K constraint.
2.1.1.1 Concept of Super Cell and Super LAC
To provide maximum integration of the femtocell network into the mobile core and macro UTRAN while at
the same time minimise the provisioning tasks the operator must undertake to achieve integration, NEC
introduces the concept of the Super Location Area and Super Cell in addition to the traditional Location
Area, Cell, UTRA Absolute RF Channel Number (UARFCN) and Scrambling Code.

Figure 1: Super LA, super cell and broadcast parameters
The Location Area and Cell ID broadcast by the FAP over the 3G air interface are the Location Area and
Cell ID that are visible to any UE monitoring that femtocell. The FAP uses a standard UMTS carrier
frequency (i.e. UARFCN) and a locally unique Scrambling Code that differentiates the femtocell from the
surrounding macro cells and adjacent femtocells.

NEC Confidential 5
th
J an. 2011
8

In the figure above, the femtocell is broadcasting using Location Area 32005, Cell ID 8035, downlink
UARFCN 10638 (corresponding to the 5MHz carrier between 2125 to 2130MHz) and a Scrambling Code
which is here represented by the letter A for the sake of simplicity.
The Super Location Area, on the other hand, is a Location Area that is configured at the Core Network
and at the IP network controller (INC). It is used to represent the core network Location Area, which
many thousands of femtocells in the INC serving area are placed into, each broadcasting their own local
Location Area Codes.
The Super Cell is a 3G Cell that is configured at the INC. Each Super Cell represents a subset of the
femtocells within the INC serving area. Therefore, a set of N Super Cells will represent all of the
femtocells in the INC serving area.
Note that, unlike in the macro UTRAN, the Super Cell is not associated with the physical area covered by
a broadcast field from a single Node B antenna. As we shall see, the Super Cell acts as a virtual cell to
represent the set of geographically dispersed femtocells that share the same UARFCN and Scrambling
Code.
2.1.1.2 UARFCN and Scrambling Code Assignment
The downlink UARFCN and Scrambling Code are air interface parameters used by the FAP to broadcast
information towards the UE. There are several considerations when selecting which values for these
parameters to use. Each femtocell must be configured as a neighbour cell to each macro UMTS
Terrestrial Radio Access Network (UTRAN) cell for cell reselection or handover, however each macro
cell neighbour list is limited in the number of entries (32 intra-frequency neighbour cells, for example). It is
not possible to consider hundreds of femtocells as neighbours to a single macro cell. Therefore we must
re-use a small set of UARFCN/Scrambling Code combinations to minimize the number of neighbour cell
entries that are allocated to the femtocell layer within each macro cell. So how many
UARFCN/Scrambling Code combinations are required? Theoretically, one is the answer, however in
this case we risk radio interference between any two femtocells that power up within range of each other.
To avoid significant radio interference in a dense population of femtocells such as might be found in an
multi-floor apartment building, there should be enough available UARFCN/Scrambling Code combinations
(e.g. 6 needed in the worst case). The actual values used for the UARFCN and Scrambling Code are
determined by the operator depending upon their frequency plan and the macro scrambling codes that
are in use.
At start-up, the FAP performs a band scanning and selects the downlink UARFCN/Scrambling Code
combination that provides the lowest interference with other femtocells and the macro network (in the
case of femtocell and microcell sharing the same frequency). Therefore, in any area where femtocells are
deployed each femtocell broadcasts using one of the six pre-defined UARFCN/Scrambling Code
combinations as shown below.

NEC Confidential 5
th
J an. 2011
9

Figure 2: Use of UARFCN/SC combinations in femtocell deployment
From the figure, each macro cell has six neighbour UARFCN/Scrambling Code combinations that
represent potential femtocell neighbours. At the Radio Network Controller (RNC), each macro cell
neighbour list will have six entries dedicated to the support of the femtocell deployment and each entry
will point to the target neighbour cell. This target neighbour cell is the Super Cell.
2.1.1.3 Mapping Super Location Area and Super Cell to Femtocell
The Super Location Area is a virtual Location Area configured both at the INC and at the MSC. Its virtual
because its never broadcast over the air. Instead, the Super Location Area exists purely to represent the
set of femtocells served by the INC to the Core Network.
Similarly, the Super Cell is a virtual 3G cell configured at the INC that is also never broadcast over the air
interface. Each Super Cell represents a subset of all the femtocells within the serving area of the INC. In
fact, each UARFCN/Scrambling Code combination in the femtocell network maps one Super Cell at each
INC. Each Super Cell has a serving area equal to that of the INC serving area and that multiple Super
Cells are therefore overlaid within the same serving area.
The diagram below shows a simplified view of the Super Cell concept through the use of two
UARFCN/Scrambling Code pairs rather than six (for simplicity).

Figure 3: Mapping femtocells to Super Cells
From the diagram, one set of femtocells uses Scrambling Code A and the other, Scrambling Code B. At
the INC, a single Super Location Area (LAC = 1000) is provisioned. Two Super Cells are also provisioned
within the Super Location Area with Cell IDs 100 and 101.
Super Cell 100 represents all femtocells that selected Scrambling Code A upon startup, and Super Cell
101 represents all femtocells that selected Scrambling Code B. In the neighbour tables within the macro
RNC, each macro cell is provisioned with two femtocell neighbours with our configured Super Cells as
the target neighbour cells.
Note that for each macro cell within the serving area of the INC, the neighbour table entries are identical:
each UARFCN/Scrambling Code combination will always point to the same target Super Cell. This
minimizes the operational overhead required to determine and populate neighbour cell table entries in the
macro network. The diagram below shows how we can imagine the overlay of each Super Cell in the
geographical area of the serving INC.

NEC Confidential 5
th
J an. 2011
10

Figure 4: Super Cells overlay the macro network concurrently
The diagram below shows that a single INC is deployed to cover the serving area of multiple RNCs. In
this case, the INC and MSC are configured in the same way as in the previous example: two Super Cells
representing the two UARFCN/Scrambling Code combinations across the network. Each RNC in the
macro network is configured with these two Super Cells as neighbour table entries to each macro cell.

Figure 5: Mapping femtocells to Super Cells multiple RNCs, single INC
The diagram below extends femtocell deployment into a more generic case where there is a second INC
and the first INC serves the same area as RNC 1, and the second INC serves the same area as RNC 3 &
5. In this case, the macro cells served by RNC 1 are configured with neighbours of Super Cells 100 and
101 as before. However, the macro cells served by RNC 3 & 5 are configured with two new Super Cells
(200 and 201) as their neighbours, both of which are provisioned on the second INC. A new Super
Location Area (1001) is defined on the second INC (RNC 4) together with the two new Super Cells.

NEC Confidential 5
th
J an. 2011
11

Figure 6: Mapping femtocells to Super Cells multiple RNCs, multiple INCs
2.1.1.4 Broadcast Location Area Code for Femtocell
The femtocell allocates itself both UARFCN & Scrambling Code parameters from a pre-defined set of
values. During registration with the INC, the macro location of the femtocell (the strongest macrocells
LAC and Cell ID from an air interface scan), the selected UARFCN and the Scrambling Code are
provided in the registration request message from the femtocell. At the INC, these values are used to map
the femtocell into the correct Super Location Area and Super Cell for representation to the Core Network
and the Super LAC/Cell ID information is returned to the femtocell in the registration accept message.
There are several very specific requirements that govern the use of Location Areas for broadcast at the
femtocell. These are:
- The Location Area must be distinct from the Location Areass used in the macro 3G UTRAN. This
is to ensure that the UE sends a Location Area Update request when the UE camps on to the
femtocell. The Location Area Update request initiates the UE to INC registration process.
- The Location Area must be locally unique. No two adjacent femtocells may have the same
Location Area to ensure that a Location Area Update is performed when the UE camps on as
described above.
- The potential set of Location Areas that can be used at the femtocell must be large enough to
minimize the risk of the UE seeing two femtocells with the same Location Area Code during
normal daily use (for UE Location Area barring purposes).
The first point is simple to address: the operator should preferably set aside a reserved range of Location
Areas that are for femtocell use only. This immediately avoids the possibility that the same Location Area
Code is present in both macro and femtocell environments. The reserved range should be as large as
possible to minimize the amount of LAC re-use in the femtocell network.
To assure locally unique Location Areas, the block of reserved LACs may be split and allocated per
UARFCN/Scrambling Code combination. In this case, two adjacent femtocells will not transmit with the
same Scrambling Code, and that is enough to guarantee that the Location Area selection will also be
locally unique.
The third point requires some additional explanation. In a network where only a subset of UEs are
authorized to access a specific femtocell (i.e. closed access model) there must be a method by which un-

NEC Confidential 5
th
J an. 2011
12

authorized UEs can be prevented from camping on the FAP. Such a method must be compatible with
existing 3G handsets and the most suited option is to use Location Area barring.
Location Area barring is a method that forces the UE not to camp on a cell that is located within a
specified Location Area. When the FAP has determined that a UE is not authorized for access, the FAP
sends to the UE location update reject message with the cause "Location Area not allowed". The UE will
add this Location Area Code as broadcast by the FAP into a list of forbidden location areas for regional
provision of service and on longer attempts to camp on any cell within the same Location Area. The list
shall be reset when the UE is switched off or when the Subscriber Identity Module (SIM) is removed, and
periodically (with period in the range 12 to 24 hours).
Therefore, it is important that a single UE should not be accidentally barred from a FAP for which it is not
authorized and which also broadcasts the same Location Area as that of the UEs home femtocell. If
such a case were to occur, the UE would no longer camp on the home FAP (it would consider itself
barred). The simplest way, in which this can be achieved is to allocate a large enough range of Location
Areas to the femtocell network to minimize the probability that such an event will occur.
2.1.1.5 Broadcast Cell Identity for Femtocell
The femtocell 3G Cell Identity can be drawn from a set-aside range of Cell IDs (if unique Cell IDs are
required for the macro network). The only requirement for the allocation of a Cell ID to the femtocell is that
it is locally unique.
The diagram below shows a typical broadcast scenario for the femtocell network. Each femtocell is
broadcasting a Location Area and Cell ID that are determined locally by the FAP, and the femtocell
provides the translation service between the FAP broadcast parameters and the INC Super Cell values.

Figure 7: Use of the femtocell Location Area and Cell Identity
For example, the FAP identified with LAC = 7002, Cell ID = 2002, UARFCN = 10638 and Scrambling
Code = B is represented to the Core Network as being located within the Super Cell LAC = 1000, Cell ID
= 101.
The macro UTRAN cell where the FAP is located will be configured to show the neighbour cell UARFCN
= 10638 and Scrambling Code = B pointing to the target Cell ID = 101 on RNC 2 (the INC). The broadcast
Cell ID of Femtocells is not populated by the macrocells in their neighbour cell list. Instead, the Super Cell
ID of the FAPs is pre-configured in the macrocells neighbour cell list.
2.1.1.6 Impact of using Super Location Area and Cells
The implementation of Super Location Area (SLA), Super Routing Area (SRA), and Super cells solves the
critical technical and operational problems associated with supporting large numbers of FAP on the

NEC Confidential 5
th
J an. 2011
13

legacy core network. However, this technique changes the way the core network and the UE work with
LAs and RAs.
The LAC/RAC information sent to the UE is different from that sent to the core network. As a result, the
UE stores the local LAC/RAC on the UEs SIM upon successful Location Area Update (LAU)/Routing
Area Update (RAU). Upon rove-out to the macro network, the UE will trigger location update and routing
area update using these local values for LAC and RAC. The core network does not have any information
about this local LAC/RAC value since the MSC/SGSN is only aware of the Super LAC/RAC for that UE.
This results in the following functional Impacts:
- UEs in idle mode with existing Packet Data Protocol (PDP) sessions will lose those PDP sessions
and will need to re-establish those PDP sessions after rove-out.
- When the UE performs a LAU or RAU after rove-out using a Temporary Mobile Subscriber
Identity (TMSI) or Packet-Temporary Mobile Subscriber Identity (P-TMSI), the MSC/SGSN will
need to identify the UE with an Identity Request message to the UE.
However, the benefits of the Super LA/RA and Super Cell concepts far outweigh these functional trade-
offs.
2.1.2 Cell Reselection
2.1.2.1 Measurement Trigger
The FAP is designed to be strictly compliant with standard Third Generation Partnership Project (3GPP)
specifications so far as Idle Mode Cell Reselection is concerned. This requirement is critical in ensuring
that interoperability with the Femtocell network demands no modifications to standard 3G Handsets. The
FAP can be operated on a carrier frequency which is the same as or different from the Macrocell carrier.
The FAP can be installed in an area where there is no 3G coverage, but where 2G coverage exists.
Hence the femtocell design supports the following Cell Reselection scenarios:
- Intra Frequency Cell Reselection,
- Inter Frequency Cell Reselection and
- Inter Radio Access Technology (RAT) Cell Reselection.
There are certain Cell Biasing Parameters defined in relevant 3GPP specification which can be used to
encourage the Cell Reselection to the FAP functionality. The FAP requires these parameter values to be
set in the Macro Layer that will ensure measurements of FAP cells to be triggered in the UE, and to bias
the UE towards attaching to the FAP within its service area. The parameters required depend upon the
specifics of the FAP deployment (for example whether the FAP shares the same UARFCN as the
Macro Layer). Typically the parameters will include:
- S
intrasearch
(SIB 3/4)
- S
intersearch
(SIB 3/4)
- Q
hyst
(SIB3/4)
- Q
offset
(SIB 11/12)
The measurement rules that apply to Idle, URA_PCH and CELL_PCH modes are defined in section
5.2.6.1.1 in 3GPP TS 25.304. It is apparent from the procedure in section 5.2.6.1.1 in 3GPP TS 25.304
that cell reselection measurement can only be triggered when S
qual

is less than or equal to the threshold
S
intrasearch
or S
intersearch
. S
qual
in this case is a measurement on the serving cell, based on the common pilot
signal to interference ratio (i.e. CPICH Ec/Io) carried out by the UE.

NEC Confidential 5
th
J an. 2011
14

2.1.2.2 Inter-frequency measurement
In the event that the Femtocell is using a different carrier as the Macro Layer, what mechanism will
ensure that S
qual
will go below the threshold (S
intersearch
) to trigger an inter-frequency measurement, given
that the Macro Layer coverage within the building would be of a high quality and/or signal level? From the
indoor Macro Layer measurements carried out within a residential property located relatively close to a
nodeB, it is found that due to local shadowing within the building the CPICH Ec/Io of the dominant
scrambling code shows sufficient variance to result in a high probability of triggering the inter-frequency
measurement. It is assumed that S
intersearch
will be a CPICH Ec/Io range of -6dB, -8dB and -10dB for
triggering. The measurement analysis suggests for these threshold values that it would take
approximately 5, 13 or 26 seconds respectively to trigger the inter frequency measurement. During that
time period the user would continue to operate on the Macro Layer within the home protected by the
adjacent channel filtering in the Femtocell. There may be a delay in cell reselection, this should not result
in call drops as the Macro Layer coverage is essentially unaffected by the Femtocell in the home due to
the adjacent channel protection. It has been suggested by a large operator that a value of -10dB might be
used as this is the S
qual
value between intra-frequency (-6dB) and inter-RAT (-14dB).
2.1.2.3 Intra-frequency measurement
Similarly in the event that the femtocell is using the same carrier as the Macro Layer, what mechanism
will ensure that S
qual
will go below the threshold (S
intrasearch
) to trigger an intra-frequency measurement to
result in the femto cell selection when the mobile moves into the building? It is suggested that this will
occur through the femtocell Radio Resource Management (RRM) algorithms that adaptively adjusts the
maximum (total) downlink transmit power to ensure femtocell Common Pilot Channel (CPICH) dominance
within the in-building coverage area.
At installation the femtocell uses the Listen Mode to select a suitable carrier/scrambling code and
populate the neighbour cell list. If the femtocell selects a carrier that is being used by the Macro Layer, as
it considers it to be the most optimal local choice, it will then set its initial total down link transmit power to
attempt to achieve its CPICH Ec/Io dominance within the in-building coverage area, based on the
detected Macro Layer Received Signal code Power (RSCP) level and an estimated indoor path loss (a
parameter defined in the Management System typically 80dB). Since this initial estimate can not
guarantee for ever Femto Cell CPICH Ec/Io dominance within the coverage area, consequently the
Femtocell utilizes ongoing UE CPICH Ec/Io measurements to adjust the total down link transmit power
level until 90% of all Femtocell CPICH Ec/Io measurements exceed that of the Macro Layer by a nominal
amount.
It is assumed that when the CPICH Ec/Io of the serving cell falls below -6dB the intra-frequency
measurement is triggered. S
intrasearch
is typically set at a point which is a trade-off between number of UE
measurements and subsequent cell search procedure (which reduces battery life) versus time taken to
ensure that the best cell is selected. It is suggested that intra-frequency measurement should be
triggered when the user enters the residence.
2.1.2.4 Rove-In
FAP UARFCN and Scrambling Code are in the Macrocells neighbour list. The FAP Cell biasing
parameters (like S
search
, Q
hyst
, Q
offset
etc) encourage Rove-In, even when the FAP has a lower quality value
and Rx level than the macro cell. Following the cell reselection measurements, the UE in idle and
connected mode would then select a Femtocell if it is deemed to be the best based on a ranking algorithm
described in section 5.2.6.1.4 in 3GPP TS 25.304. PLMN
Femto
can equal PLMN
Macro
or PLMN
Femto
is
flagged as an Equivalent Public Land Mobile Network (PLMN) to PLMN
Macro.
After deciding on Rove-In,
based on the cell reselection measurements and criteria, the UE shall perform Location Update procedure
to register with the FAP due to the different LAC/RAC broadcast by the FAP:

NEC Confidential 5
th
J an. 2011
15

Figure 8 Cell Reselection to Femtocell Procedure
The Rove-in feature for macro 3G to femto 3G cell selection and macro 2G to femto 3G cell selection are
supported in Release 1.0.
2.1.2.5 Rove-Out
Surrounding macrocell UARFCN / Scrambling Codes are populated in the FAP neighbour list. The FAP
Cell biasing parameters (like S
search
, Q
hyst
, Q
offset
etc) discourage Rove-Out (the FAP is sticky). Following
the cell reselection measurements, the UE would then select a Macrocell if it is deemed to be the best
based on a ranking algorithm described in section 5.2.6.1.4 in 3GPP TS 25.304. After deciding on Rove-
Out, based on the cell reselection measurements and criteria, the UE shall perform Location Update
procedure to register with the Macro network due to different LAC/RAC between the FAP and surrounding
Macrocell.

NEC Confidential 5
th
J an. 2011
16

Figure 9 Cell Reselection to Macro Network Procedure
Note: Identity Request / Response for IMSI-1 is additionally triggered on reselection to re-acquire UE
Identity because the core network does not have any information about the broadcast LAC value (LAI-y
n
)
and the MSC/SGSN is only aware of the Super LAC/RAC for the UE.
The Rove-out feature for macro 3G to Femto 3G cell selection and macro 3G to Femto 2G cell selection
is available in Release 1.0.

2.1.2.6 Configuration for Cell Selection/Reselection
Cell Selection
Cell selection in idle mode is strictly compliant to 3GPP TS 25.304 and 3GPP TS 25.331 on cell ranking
algorithm. The system information parameters for all cell selection and reselection in idle and dedicated
modes are operator configurable.
The UE will read the values of Q
qualmin
, Q
rxlevmin
and Maximum allowed Uplink (UL) Transmit (TX) power
broadcast in the FAPs System Information Block Type 3 (SIB 3). The UE will then use these values to
determine whether the FAP cell is a suitable cell as defined in section 5.2.3 Cell Selection Process in
3GPP TS 25.304 , i.e.,
S
qual
= Measured CPICH Ec/No - Q
qualmin
> 0
S
rxlev
= Measured CPICH RSCP - Q
rxlevmin
- P
compensation
> 0
Where P
compensation
= max(Maximum allowed UL TX power Maximum Radio Frequency (RF) output
power of the UE, 0).
The System Information parameters for cell selection in idle mode (SIB 3) broadcast by the FAP can be
set in two ways:
- AP-MS provides (configurable) static parameters values
- AP-MS provides the algorithms (configurable) to be used by the FAP to set the parameters
according to measurements of surrounding cells/radio environment.

NEC Confidential 5
th
J an. 2011
17

Finer control of the UE behaviour regarding decision to camp/stay camped on the FAP is best achieved
through tuning of the cell reselection parameters. The cell selection parameters should be set mainly with
the intention that the UE only camps/stay camped on the FAP if its quality is good enough not to lead to
degradation in the user experience, e.g. call setup failures due to radio link failure.
Cell Reselection
Cell reselection between the 3G macro network and the FAP (and vice versa) is strictly compliant to
3GPP TS 25.304 .
When the UE is camped on a UTRAN macrocell in idle mode:
- It will perform cell reselection procedures in idle mode as defined in 3GPP TS 25.304 in
accordance to the cell reselection parameters values broadcast by that cell in SIB 3 and SIB 11.
- The FAP will be seen by the UE as a neighbour cell because it will be using one of the <UARFCN,
Scrambling Code> pairs listed in the Macro cells neighbour list for idle mode in SIB 11.
When the UE is camped on a FAP in idle mode:
accordance to the cell reselection parameters values broadcast by that FAP in SIB 3 and SIB 11.
- Surrounding UTRAN Macro cells will be seen by the UE as neighbour cells because their
<UARFCN, Scrambling Code> will be listed in the FAPs neighbour list for idle mode in SIB 11.
Cell reselection from the FAP to the 2G Macro network is strictly compliant to 3GPP TS 25.304. When the
UE is camped on a FAP in idle mode:
accordance to the cell selection/ reselection parameters values broadcast by that FAP in SIB 3 and
SIB 11.
- Surrounding GERAN Macro cells will be seen by the UE as neighbour cells because their <ARFCN,
BSIC> will be listed in the FAPs neighbour list for idle mode in SIB 11.
Cell reselection from the 2G macro network to the FAP is strictly compliant to 3GPP TS 45.008 and3GPP
TS 43.022. When the UE is camped on GERAN macrocell in idle mode:
- It will perform cell reselection procedures in idle mode as defined in 3GPP TS 45.008 and 3GPP
TS 43.022 in accordance to the cell reselection parameter values broadcast by that cell in System
information type 2quarter.
- The FAP will be seen by the UE as a UTRAN neighbour cell because it will be using one of the
<UARFCN, Scrambling Code> pairs listed in the Macro cells 3G neighbour list for idle mode in SI
2quater.

Intra Frequency Cell Reselection
When the UE is camped on a 3G cell (UTRAN Macro cell or a FAP), cell reselection to a surrounding
UTRAN cell is only possible if the serving cell lists the <UARFCN, Scrambling code> pair used by the
surrounding cell in its neighbour list broadcast in SIB 11.
When the UE is camped on the FAP, it will rank the serving cell (FAP) and surrounding macrocells
included in its neighbour list as defined in 3GPP TS 25.304 in accordance with the cell selection /
reselection parameters broadcast by the FAP in SIB 3 and SIB 11 in exactly the same ways as when it is
camped on a UTRAN macrocell.
Unless the operator desires to set fixed parameter settings via the AP-MS, the FAP self-configures the
cell reselection parameters it broadcasts in SIB 3 and SIB 11 such as to minimise ping-pong between the
FAP and the macro layer. It achieves this by dynamically taking into account the values of the cell

NEC Confidential 5
th
J an. 2011
18

reselection parameters broadcast by the UTRAN macrocells that list the FAP as a neighbour cell (and are
in turn listed by the FAP as neighbour cells).
More specifically, FAP will read the cell reselection parameters (e.g. Qhyst1,s, Qhyst2,s, Qoffset1s,n and
Qoffset2s,n) broadcasted by each of these cells. The FAP will then compute the bias that is being applied
by each of these UTRAN cells relative to the FAP. The computation will depend on the cell reselection
quantity used in that cell, i.e. CPICH RSCP or Ec/No. It is assumed that all surrounding cells will be using
the same quantity.
If the quantity used is CPICH RSCP then the bias is computed as
- Bias (3G FAP) = Qhyst1,s + Qoffset1s,n
If the quantity used is CPICH Ec/No then the bias is computed as
- Bias (3G FAP) = Qhyst2,s + Qoffset2s,n
The FAP will then set its cell reselection parameters (Qhyst1,s , Qhyst2,s , Qoffset1s,n , Qoffset2s,n)
broadcast in SIB 3 and SIB 11 to generate a bias in the reverse direction, Bias (FAP-> 3G), that leads a
well defined hysteresis between reselecting from the Macro Layer to the FAP and from the FAP to the
Macro Layer.
- Bias (FAP 3G) + Bias (3G FAP) = hysteresis (dB)
This hysteresis prevents the ping pong effect between the Macro network and the FAP as a user moves
around close to its FAP.
Note the following:
- There is no need for the operator to fine-tune cell reselection parameters broadcast in its macro
cells in order to preclude ping pong effects as the FAP can dynamically adapt to the parameter
values broadcast by the macro cells
- The logic/algorithm used by the FAP to self-configure the cell reselection parameters it broadcasts
is provisioned by the AP-MS and can be updated by the operator.
Inter Frequency Cell Reselection
If the FAP is operating on a UARFCN different from that of the UEs serving cell, then the UE will consider
the FAP as part of its inter-frequency neighbour list. In order for the FAP to be seen as a suitable
candidate for cell reselection, its CPICH level must be high enough to meet the S criteria as defined by
Q
qualmin
and Q
rxlevmin
which are broadcast in the SIB 3.
In addition, the serving cell may inform the UE that it need not perform measurement on neighbour cells
as long as the serving cell Ec/No is higher than a certain value. For inter-frequency neighbours, this
behaviour is indicated by broadcasting a value for the S
intersearch
parameter (SIB 3 of serving cell). If
theparameter is included, then as long as measured serving cell
CPICH Ec/No > Q
qualmin
(SIB 3 of serving cell) + S
intersearch

the UE is not obliged to initiate measurements on inter-frequency neighbours.
Thus the cell reselection behaviour of a UE while camped on a cell is completely under the control of the
serving cell as it is ruled by the cell reselection parameters broadcast by that cell in SIB 3 and SIB 11.
Hence the only way the UE will cell reselect from a macro cell to the FAP when the signal strength/quality
(RSCP or Ec/No according to the ranking quantity parameter chosen by the macro layer) of the macro cell
exceeds that of the FAP is that if the macro cell broadcasts cell reselection parameters that bias the cell
ranking algorithm towards the FAP via a suitability negative values Qoffset1s,n and Qoffset2s,n.
Due to the typical nature of the deployment of the FAP and in close proximity of the UE, it is judged to be
very unlikely that such a strong macro dominance will occur and a small (global i.e. not macro cell

NEC Confidential 5
th
J an. 2011
19

specific) positive bias will be enough to guarantee that the UE will reselect to the FAP when in its close
proximity.
When the UE is camped on the FAP, its ranking/reselection behaviour is completely under the control of
the FAP (via the settings of the parameters in SIB 3/11). Thus regardless of the value of the bias (toward
the FAP) used by the macro layer cells, the FAP will always set its own cell ranking parameters (SIB 3/11
of the FAP) to values that guarantee an adequate hysteresis between the MacroFAP reselection
relative to FAPMacro reselection. Thus the FAP can ensure that the UE does not ping pong back and
forth between macro Layer and FAP without the operator having to fine-tune cell reselection parameters
broadcast in its macro Layer.

2.1.3 Handover
2.1.3.1 Hand-Out (Handover from 3G Femtocell to 3G Macrocell)
In NEC femtocell Solution, the hand-out functionality is perceived as Combined Hard Handover and
Serving Radio Network Subsystem (SRNS) Relocation procedure as described in 3GPP specification. In
3GPP this procedure is suggested to be used for handover between two cells each belonging to a
different Radio Network Controller (RNC) in the absence of Iur (i.e. RNSAP Interface) connection
between those RNCs.
We outline the following key functional procedures to enable handover from femtocell to macrocell
scenarios:
- Population of candidate macrocell information in femtocell Neighbour Cell List
- Configuration of required measurement events to be performed and reported by UE
- Determine the Cell-Identity of Neighbour macrocell(s) to uniquely identify the target cell.
The FAP is initialised, it scans the operators carrier frequencies (this information is provided by the AP-
MS to discover the UARFCN and Scrambling Code combination of surrounding UTRAN macro cells and
use this information to build an intra-frequency and inter-frequency neighbour macrocell list. The FAP will
decode each of these detected cells Broadcast Information (SIB 3) to obtain its 28-bit Cell ID (along with
Location Area Identifier (LAI) and Routing Area Identifier (RAI)). The Cell ID uniquely identifies the cell in
the PLMN, which is needed during handover to identify a Target cell. Therefore the FAP will know the 28-
bit cell ID of each intra/inter-frequency macro cell it lists in its neighbour lists. The FAP also decodes the
necessary Broadcast Information (SIB 11) of strongest Macrocell to obtain its Neighbour Cell List, which
helps to derive other neighbour cells.
The FAP collates the detected and derived neighbour cells in the following manner:
- Divide the cells into two categories: Intra and Inter Frequency Neighbour Cells lists.
- In each list, sort the cells starting with stronger signal strength detected by Scanning the Carrier
and Scrambling Codes and for which FAP has decoded the corresponding Cell Identities.
- Append additional Neighbour Cells as obtained from Broadcast Information (i.e. from Macro Cell
Neighbour Cell List of strongest Macrocell) for which Cell Identities are decoded successfully.
- Append additional Neighbour Cells as obtained from Broadcast Information (i.e. from Macro Cell
Neighbour Cell List of strongest Macro Cell) for which Cell Identities could not be decoded.
- The Neighbour Cell List size is limited to 10 hence only first 10 strongest cells added in the list and
rest are discarded.
In order to enable handover from the FAP to surrounding UTRAN macro cells, the FAP will provide this
neighbour macro cell list to the UE and configure periodic intra/inter-frequency measurements in the UE
while in call.

NEC Confidential 5
th
J an. 2011
20

The FAP will contain a handover decision algorithm provisioned by the AP-MS, taking into account the UE
measurement reports.
Since the FAP is responsible for initiating the handover request towards the INC, the FAP is in full control
of when the UE moves from the FAP to the macro layer, and so can preclude the UE from being handed
over back and forth between the FAP and the macro layer.
When the handover decision algorithm determines the need to perform the handover procedure it will
signal this to the INC via a HANDOVER REQUEST message. The hand-out procedure involves the UE,
serving FAP, INC, Core Network and Target RNC as shown below.

FAP UE
INC
1. Uu - Measurement Report
Target
RNC
3. Relocation Required
Ongoing Voice Call
2. GA-CSR RELOCATION REQUIRED
8. Uu Physical Channel
Reconfiguration
14. Voice Traffic
CN
4. Relocation Request
5. Relocation Request Ack
6. Relocation Command
7. GA-CSR RELOCATION COMMAND
10.. Relocation Detect
9. Uu UL Synchronization
11.. Voice
12. Uu Physical Channel Reconfiguration Complete
13.. Relocation Complete
15. Iu Release Command
16. GA-CSR RELEASE
17. GA-CSR RELEASE COMPLETE
19. GA-RC DEREGISTER (UE)
18. Iu Release Complete

Figure 10 Hand-Out procedure
The HANDOVER REQUEST message will contain the 28-bit cell Identity read from SIB 3 of the target
macro cell.
In networks where the cell identity broadcast in SIB 3, the Cell Id is the same as used for signalling
between UTRAN Network elements and the INC simply maps the received Cell Id to the Target Cell ID
field (in the Source RNC To Target RNC Transparent Container) and uses the first 12 bits to set the
Target RNC ID in the RANAP: Iu RELOCATION REQUIRED message.Otherwise the INC will have to
access a database (using the Cell Identity in SIB 3) to discover the Target cell ID and RNC ID of the RNC
that controls the target cell.
The INC will then act as the Source RNC and generate the Iu RELOCATION REQUIRED message. The
Core Network will use the Target RNC ID to discover the Target RNC which will use the Target cell ID to
contact the target Node B and set up the new radio link.
Reception of the Layer 3 confirmation message (e.g. PHY CH RECONFIGURATION COMPLETE) by the
target RNC triggers the target Radio Network Controller (RNC) to send Iu RELOCATION COMP to the
Core Network, which will result in the release of the Iu-CS to the INC. The INC will then release the
Generic Access Circuited-Switched Resource (GA-CSR) connection and the FAP will de-register the UE
from the INC.

NEC Confidential 5
th
J an. 2011
21

2.1.3.2 Hand-Out (Handover from 3G Femtocell to 2G Macrocell)
To support handover from 3G femtocell to 2G macrocell, the FAP needs to be populated the list of 2G
neighbour cells. The 2G neighbour cells are collated in the following ways:
- Obtain the 2G Neighbour Cell list as broadcast by 3G Macro Cell (i.e. derived from 2G Neighbour
Cell List of the 3G macrocell).
- Decode the Broadcast information of these 2G cells to obtain their Cell Global Identities (CGI).
The FAP then adds those 2G cells belonging to the Operator (based on PLMN Id) in the inter-RAT
Neighbour Cell list in the form of Absolute Radio Frequency Channel Number (ARFCN), Base Station
Identity Code (BSIC) and CGI. Alternatively, the AP-MS can provide CGI information to FAP based on
FAP geographical position.
In addition, the FAP uses compressed mode to enable UE measurements on GSM neighbour cells. This
is a necessary condition for supporting handover to GSM for UEs that do not have a dual 2G/3G receiver.

FAP UE
INC
1. Uu - Measurement Report
Target
BSC
3. Relocation Required
Ongoing Voice Call
2. GA-CSR RELOCATION REQUIRED
8. Uu Handover from UTRAN
(DTAP HO Command)
15. Voice Traffic
CN
4. Handover Request
5. Handover Request Ack 6. Relocation Command
(DTAP HO Command)
7. GA-CSR RELOCATION COMMAND
(DTAP HO Command)
10.. Handover Detect
9. Um Handover Access
11.. Voice
13. Um Handover Complete
14.. Handover Complete
16. Iu Release Command
17. GA-CSR RELEASE
18. GA-CSR RELEASE COMPLETE
20. GA-RC DEREGISTER (UE)
19. Iu Release Complete
12. Um Physical Information

Figure 11 Inter-RAT Hand-Out
2.1.3.3 Supported Hand-Out Scenario
The hand-out feature for
- Inter frequency handover from femto 3G to macro 3G - Circuit Switched case
- Intra frequency handover Femto 3G to macro 3G - Circuit Switched case
- Voice call with Femtocell it is handed over to 2G macro cell
- Video call handout to macro 3G

NEC Confidential 5
th
J an. 2011
22

- Multi-RAB handout (voice only) with Packet Switched release (femto 3G to macro 2G/3G)
is available in Release 1.0.
- PS call handout to macro 3G
- Multi-RAB handout to macro 3G (CS+PS)
will be available in Release 2.0.

2.2 Radio Resource Management
Radio Resource Management (RRM) algorithms encompass strategies and techniques for efficient
utilisation of limited radio resources and control of co-channel interference. RRM is needed to guarantee
QoS, to maintain the planned coverage area, and to offer high capacity. Typically RRM controls
parameters such as transmit powers, handover criteria, channel allocation, admission criteria, etc. The
family of RRM algorithms can be divided into handover control, power control, admission control, load
control, and packet scheduling functionalities. Power control is needed to keep the interference level at
minimum in the radio interface and to provide the required QoS for radio bearers. Handovers are needed
to handle the mobility of the UEs across cell boundaries. Other RRM algorithms like admission control,
load control and packet scheduling are required to guarantee the QoS and to maximise the system
throughput with a mix of different bit rates, services and quality requirements.
The FAP solution is for low cost residential/small office usage with a supporting AP-MS. Since the FAP is
targeted as a consumer/mass product similar to Wireless LAN, it is expected that the user would be
responsible for the deployment of the FAP, which is a departure from the traditional Radio Access
Network design typically carried out by the operator. This difference could result in higher than planned
levels of interference, which would impact on network coverage and QoS. Thus, the FAP deployment by
subscribers in an ad hoc manner requires a different approach to RRM algorithms.
A network of FAPs is remotely managed by the operator through the AP-MS, which provides all the
functionality required to deploy and manage the FAPs. In addition, the FAP network is intended to coexist
with the Macro network with the possibility of sharing radio resources such as carrier frequency, which
presents a whole new set of challenges.
This section describes the details of the RRM algorithms that are employed within the FAP to minimise
the effects on neighbour macrocells and FAPs while maintaining a suitable local QoS.
2.2.1 Interference Management
Two important principles in interference management guide RRM algorithms in the FAP implementation
that the FAP utilises to minimise its impact on the coverage and QoS of the Macrocells and surrounding
Femtocells. These are as follows:
- The macro network has the highest priority in terms of coverage and QoS. Therefore, the FAP
should always tend towards minimising any impacts on the macro network, even at the cost of its
degraded local QoS.
- FAP should always tend towards minimising the impacts on surrounding femtocells, even at the
cost of its degraded local QoS.
Initial Maximum Uplink and Downlink Transmit Power
With the carrier and scrambling code selected, the FAP will calculate the initial maximum uplink and total
downlink transmit powers using the proprietary FAP initial power settings algorithm.
At Cold Start, the FAP selects the carrier that exhibits the lowest interference level, where the
interference is a sum of all the CPICH RSCP powers of the surrounding neighbour cells. Following the

NEC Confidential 5
th
J an. 2011
23

selection of a carrier, the FAP then selects an unused scrambling code (from the list provided by the AP-
MS). In the event that all FAP scrambling codes are used, the scrambling code with the lowest RSCP is
selected. During this process the FAP also detects error conditions. For example, if the local interference
level exceeds a particular level (typically -65dBm within the home, which equates to approximately -
50dBm external to the building) for the best carrier, it is concluded that an error condition has occurred
and the FAP would not transmit. An error message would be reported to the AP-MS and the user
informed.
By decoding the Broadcast Channel (BCH) of the macro and FAP neighbour cells, the CPICH Tx power is
known and the path loss from the neighbour macro Node Bs and neighbour FAPs to the FAP can be
calculated. The initial UL TX power is calculated using these path losses in order to minimise the noise
rise at the macro Node B. The initial maximum downlink transmit power is determined by the highest
macrocell CPICH RSCP for the selected carrier and a nominal indoor path loss. Both the maximum
transmit powers are within the bounds set by the AP-MS.
Dynamic Adaptation of Maximum Transmit Power
The power adaptation algorithms are used afterwards for setting the ongoing maximum total downlink and
uplink transmit powers. In addition, due to the dynamic range of the FAP systems transmission powers
operating outside the 3GGP specified range, the FAP power control also modifies some of the information
elements of the system information blocks used in the uplink open-loop power control equations.
In idle periods, the FAP continues to monitor both the downlink and uplink and builds a log of the local
environment: e.g.: surrounding FAP and macrocell CPICH RSCP levels, number of surrounding Macro
Node Bs and FAPs and carrier/scrambling code utilisation. The FAPs initial selection of maximum Tx
power, carrier frequency and scrambling code would be recalculated if there is any significant change in
the radio environment as determined by the log such as a large increase in downlink interference or a
large uplink noise rise.
Once in a call the FAP also uses UE ongoing neighbour cell measurements to update the path loss
estimate, and thereby controls the maximum FAP UE transmit power (maximum UL transmit power).
Histograms (with varying time constants) of ongoing measurements of connected UEs are used to keep
the maximum FAP transmit power to a minimum.
During data calls, FAP uses transmitted power logs and interference measurements to dynamically adjust
its data rates to cope with unwanted interference such as an unregistered UE coming within the FAP
coverage area.
When making a change to the maximum uplink and downlink transmit power levels, the relevant SIBs are
updated and broadcast on the BCH. In the downlink, the proportion of power allocated to specific
Dedicated Channels (DCHs) and control channels remains unchanged and only the total power level is
adjusted.
2.2.2 Power Control
Power control is crucial for Code Division Multiple Access (CDMA) systems for the following reasons. On
the uplink, excessive power (i.e. noise rise) from a single UE can significantly impact the cell coverage
and cause dropped calls at the cell boundary. On the downlink, the system capacity is directly determined
by the required code power for each transmission. It is therefore essential to keep the transmission
powers at a minimum level while ensuring adequate signal quality at the receiving end. The group of
functions used to achieve this are: open-loop power control, inner-loop (or fast) Power Control (PC) and
outer-loop PC in both uplink and downlink directions. Slow PC is also applied to the downlink common
channels. Open-loop PC is responsible for setting the initial uplink and downlink transmission powers
when a UE is accessing the network. The outer-loop PC estimates the received quality and adjusts the
target Signal to Interference Ratio (SIR) for the fast closed-loop PC so that the required quality is provided.

NEC Confidential 5
th
J an. 2011
24

2.2.2.1 Open-Loop Power Control
Significant correlation exists between the average path loss of the uplink and downlink frequency bands
over time and hence calculations made on the downlink are often applied to the uplink when setting the
initial power values.
In the downlink, the CPICH pilot, which is usually a percentage (typically 10%) of the total maximum
transmit power, defines the coverage area for the FAP and in particular the QoS within the coverage area.
In order for macrocell UEs to camp on to the FAP, an area of CPICH dominance needs to be created.
However, excessive areas of dominance cause unnecessary levels of interference to macrocell as the
macrocell UEs, which are excluded from roaming into the FAP coverage area, need a higher proportion of
macrocell Tx power to overcome the local interference.
When calculating the initial maximum downlink and uplink transmit powers the following scenarios need to
be taken into consideration:
- The initial downlink/uplink maximum power calculations should not exceed the limits as defined by
the AP-MS.
- If Macrocell interference level exists or exceeds neighbour FAP interference, then the FAP will set
its initial downlink power based on the strongest Macrocell CPICH RSCP level. The initial uplink
power will be set based on a typical Macrocell noise floor with an additional back off margin applied
to limit the additional uplink noise rise caused by the FAP UEs to the Macrocell and surrounding
FAPs.
- In the case of the FAP selecting a carrier that is not being used, the initial downlink and uplink
power settings are set to fixed values (enough to overcome the FAP UE noise floor and FAP noise
floor respectively and allow for a nominal indoor path loss of around 95dB).
- If the path loss to the neighbour FAP is less than a certain amount, then it sets its downlink power
at the same level as the neighbour FAP. This is done by reading the BCH of the strongest
neighbour FAP. By doing this, all collocated FAPs will be slaved off the first FAP that has set its
power level from the Macrocell CPICH RSCP level. If this neighbour FAP powers down then the
FAP through the listen mode will detect this scenario and recalculate its initial maximum downlink
and uplink transmit power level. This is done for the following reasons:
- If the FAP powers up very close to another FAP operating on the same frequency, it is
unlikely the FAP will be able to measure macrocell signals on the frequency it is operating
and hence will not be able to set powers that minimise the interference to macrocells. This
also covers the neighbour cells list creation as the FAP will copy the neighbour cell list of the
other FAP.
- Two collocated FAPs with overlapping coverage areas, (e.g. when FAPs are deployed on
either side of a common wall to 2 residences) if allowed to power adapt, will likely end up
pushing their powers up to the maximum as they both try to achieve better signal quality
within their coverage areas leading to an unnecessary increase in interference.
- As a starting point for both uplink and downlink powers, a typical indoor path loss of 95dB is
allowed (this value is programmable through AP-MS). This will provide a typical coverage of
approximately 50 to 150m
2
(ITU-1238 model) depending on building materials etc.
- A key assumption in the maximum downlink power calculation is that the percentage CPICH power
allocation as part of the maximum downlink transmit power is the same for all FAPs. This is
enforced through the AP-MS across all FAPs.
Uplink Open-Loop Power Control
The uplink open-loop PC function is located both in the terminal and in the UTRAN and requires some
control parameters being broadcast in the cell and the CPICH RSCP being measured by the terminal.
Based on the calculation of the open-loop PC, the terminal sets the initial powers for the first Random
Access Channel (RACH) preamble and for the uplink Dedicated Physical Control Channel (DPCCH)

NEC Confidential 5
th
J an. 2011
25

before starting the inner-loop PC. During the random access procedure the power of the first transmitted
preamble is set at:
CI quired Re UL ce Interferen UL RSCP CPICH Power Tx CPICH
Power Initial eamble Pr
_ _ _ _ _ _
_ _
+ + =

Where CPICH_Tx_power, the transmit power of the CPICH, and UL_Required_CI, the required C/I in
uplink, are set during Radio Network Planning and the UL_interference is calculated in the Node B. All
three parameters are broadcast on the BCH. When establishing the first DPCCH, the UE starts the uplink
inner-loop PC at the power level according to:
RSCP CPICH offset Power DPCCH power Initial DPCCH _ _ _ _ _ =
Where CPICH_RSCP is measured by the terminal and the DPCCH_Power_offset is typically calculated
by the Admission Control (AC) in the RNC and provided to the UE at Radio Resource Control (RRC)
connection setup or during radio bearer or physical channel reconfiguration, as:
) ( log 10 _ _ _ _ _
10 DPDCH DPCCH
SF SIR erference int UL power Tx CPICH Offset Power DPCCH + + =

Where the SIR
DPCCH
is the initial target SIR for that particular connection and the SF
DPDCH
is the spreading
factor of the corresponding DPDCH.
FAP has been designed to operate over a wide dynamic range of transmission powers (+10dBm to -
40dBm) to cover a range of deployment of scenarios. Typically, 10% of total down link power is allocated
to CPICH, which implies CPICH transmission power will be in the range of 0dBm to -50dBm. However,
this presents a problem as 3GPP standards give the range of the CPICH transmission power broadcast
in the system information block as 50dBm to -10dBm.
To avoid excessive uplink transmission powers, the variables
ce Interferen UL_
and
Power Tx CPICH _ _

used above are adjusted as shown below:
Adjustment CPICH measured ce Interferen UL ce Interferen UL _ _ _ _ =
Where
measured ce Interferen UL _ _
is the actual measured wide band receiver noise power (RSSI)
>
=
otherwise real Power Tx CPICH Min Power Tx CPICH
Mn Power Tx CPICH real Power Tx CPICH if
Adjustment CPICH
, _ _ _ _ _ _
_ _ _ _ _ _ , 0
_

Where
dBm Min Power Tx CPICH 10 _ _ _ =
, is the lower limit of the CPICH transmission power range and
real Power Tx CPICH _ _ _
, is the actual CPICH transmission power of the FAP.
Adjustment CPICH real Power Tx CPICH Power Tx CPICH _ _ _ _ _ _ + =
The range of Uplink Interference is given as -70 to -110 dB. On occasions where the measured Uplink
Interference is at the lower limit, for example when an exclusive carrier has been assigned to the
femtocell network and there is no macrocell/femtocell activity on the carrier, there will be no headroom to
apply a bias value. However, in such a situation, as no femtocell or macrocell activity was detected,
transmitting at too high a power would not be interfering with any FAP or Node B. The variable
Adjustment CPICH _
is stored in the system database.
Downlink Open-Loop Power Control
In the downlink, the open-loop PC is used to set the initial power of the downlink channels based on the
downlink measurements reports from the UE. This function is located in both UTRAN and UE. A typical

NEC Confidential 5
th
J an. 2011
26

algorithm for calculating the initial power value of the Dedicated Physical Data Channel (DPDCH) when
the first bearer service is set up is:
( )
|
|
.
|
\
|
= PtxTotal
N E
power Tx CPICH
W
N E R
P
CPICH o c
DL o b Initial
Tx
o
/
_ _ ) / (

Where R is the user bit rate, (E
b
/N
o
)
DL
is the downlink planned E
b
/N
o
value set by Radio Network Planning
for that particular bearer service, W is the chip rate, (E
c
/N
o
)
CPICH
is reported by the UE, is the downlink
orthogonality factor, and PtxTotal is the carrier power measured at the Node B and reported to the RNC.
For a modified radio bearer the scaling is done with the new user bit rate and new downlink E
b
/N
o
.
Power Control on Downlink Common Channels
The transmission powers of the downlink common channels are determined by the network. In general
they are specified at the network set up but could be changed dynamically. Typical values for the
downlink common channels powers are as follows:
DL Common
Channel
Typical Power Level Note
P-CPICH 5-10% of the maximum cell Tx
power

P-SCH and S-
SCH
-3dB Relative to P-CPICH power
P-CCPCH -5dB Relative to P-CPICH power.
Carries BCH information
PICH -8dB Relative to P-CPICH power.
Paging Indication Channel. Np =
72.
AICH -8dB Relative to P-CPICH power
S-CCPCH -5dB Relative to P-CPICH power.
Carries FACH and PCH (Paging
Channel)
Table 1: Typical downlink common channel power levels
Please note that CPICH, P-SCH, Secondary Synchronisation Channel (S-SCH) and Primary Common
Control Physical Channel (PCCPCH) are cell specific configuration parameters which through radio
network planning set the actual cell size. It is expected in the FAP that cell size will need to be
dynamically controlled to provide the most optimal trade-off between FAP QoS while minimising the
macrocell interference impacts. To retain a fixed scaling relationship between the individual control
channels and the data channels it is expected that the maximum downlink power will be controlled to
adjust the FAP coverage as a function of macrocell CPICH RSCP. This is different to managing the
individual control channels.
2.2.2.2 Inner-Loop Power Control
Inner-loop power control relies on the feedback information at the Layer 1 from the opposite end of the
radio link. This allows the UE/Node B to adjust its transmit power based on the received SIR level at the
Node B/UE to compensate for radio channel fading. Inner-loop PC is used for dedicated channels in both
uplink and downlink. Fast PC update rate is 1.5kHz.

NEC Confidential 5
th
J an. 2011
27

The following RRC information is communicated from the RNC to the UE: Downlink Target Block Error
Rate (BLER), Uplink gain factors, Uplink TPC (power control step size), Power Control Algorithm, Uplink
Resource Management values, (DPC_MODE (Algorithm 1 or 2).
The following RRC information is communicated back from the UE to the RNC: actual BLER, CPICH
Ec/No, CPICH RSCP, path loss, traffic and UE internal measurements.
Uplink Fast Closed-Loop Power Control
With respect to this function the following brief summary is provided:
The Node B receives target SIR from the uplink outer loop PC located in the RNC and compares it
against the estimated SIR measured from the pilot symbols on the DPCCH (note DPCCH and DPDCH
are multiplexed on the I/Q channels). If measured SIR exceeds the target the Transmit Power Control
(TPC) down command is issued to UE. If the estimated SIR is less than the target SIR then the TPC up
command is issued.
The optimum PC step size depends on UE speed 1dB step size will track up to 30km/hr (55Hz fading
rate).
There are two inner-loop Power Control algorithms Algorithm 1 and Algorithm 2. Algorithm 2 provides
superior performance for fading rates lower than 3km/hr. FAP should be configured to use this algorithm.
Algorithm 2 effectively reduces the update rate to 300Hz.
In starting the uplink DPDCH, a terminal is instructed to send an uplink DPCCH PC preamble and ranges
between 0 and 7 frames as set by the Radio Network Planning. During this time period Algorithm 1 is
activated to result in fast convergence.
During the PC communication after applying the DPCCH power adjustments and gain factors the UE is
not allowed to exceed the maximum transmit power as set by the Radio Network Planning.
Downlink Fast Closed-Loop Power Control
The downlink inner-loop power control sets the power of downlink DPCH and the following brief summary
is provided:
The UE receives from the higher layers the BLER target set by the RNC for the down-link outer-loop PC
together with other control parameters. The UE estimates the downlink SIR from the pilot symbols of the
downlink DPCH. This SIR estimate is compared with the target SIR. If the estimate is higher than the
target the UE transmits the TPC command up and if the estimate is lower than the UE transmits the TPC
command down.
The relative power differences between the DPDCH and the Transport Format Combination Indicator
(TFCI), TPC and pilot fields of the downlink DPCCH are determined by the power offsets P01, P02 and
P03.

P02
O2
P02
O2
P03
O2
Dat
a
Dat
a
TP
C
TFCI
Pilot

Figure 12: Power Offsets for Improving the Downlink Signalling Quality

NEC Confidential 5
th
J an. 2011
28

The Downlink power control range is defined by the diagram below:

3dB
28dB
DL Total power
dynamic range
18dB
DL PC dynamic
range
No traffic channels activated
Minimum Code channel power
Node B Max
output power

Figure 103: Downlink Power Control Range
2.2.2.4 Outer Loop Power Control
The aim of the outer-loop power control algorithm is to maintain the quality of the communication at the
level defined by the quality requirements of the bearer service in question by producing adequate target
SIR for the inner loop power control. This operation is carried out for each DCH belonging to the same
Radio Resource Control (RRC) connection.
2.2.2.2 Adaptation of Maximum Powers
The maximum uplink and downlink power adaptation algorithms use the following concepts when
adjusting the uplink and downlink transmit power levels:
- During the first few voice calls the FAP builds a histogram of the indoor path loss, the RSCP and
the CPICH Ec/No values of the active FAP cell and neighbour cells by instructing the UE to make
the appropriate measurements.
- The objective of adapting the maximum downlink and uplink transmit power levels is to improve
the local coverage and QoS for the FAP UEs but minimise the interference effects on the
neighbour macrocells and FAPs. To do this, the basic assumptions are as follows:
- The majority of the calls made by FAP users will take place within the expected coverage
area (e.g. home and/or office) and that the maximum uplink transmit powers will be set such
that the noise rise at neighbour Macro neighbours is minimised while the downlink transmit
power levels will be set to achieve a certain CPICH Ec/No value for a predetermined number
of samples (e.g. 90%).
- Within the expected FAP coverage area there is a low probability of unregistered macrocell
UEs.
- The difference between the uplink and downlink power is then a function of the power
allocation between the different downlink channels (i.e. proportion of downlink power is
allocated to control channels, High Speed Downlink Packet Access (HSDPA)) which limits
the power available to a particular DCH.
- The macrocell uplink noise rise can be limited by using the knowledge of the path loss
spread between the FAP UEs that are roaming within the expected FAP coverage area and
the neighbour macrocell Node Bs to set a limit on the maximum allowed FAP UE Tx power.
Clearly the macro layer to FAP UE path losses measured will be different from the initial path
loss calculations made by the FAP in Downlink Listen Monitor Mode. Measurements will
typically be taken from the FAP UE when in CELL DCH.
- From measurements conducted, it assumed that the typical indoor propagation losses are
within the range of 60 to 95dB.

NEC Confidential 5
th
J an. 2011
29

- The maximum uplink and downlink transmit powers may converge to different values. The
maximum uplink transmit power may be larger than the downlink transmit power allocated to
that particular FAP UE/DCH. The extra uplink power headroom will accommodate the
situation when unregistered users come into the FAP coverage area hence leading to the
FAP-UE requiring more power to overcome the resulting interference scenario.
- When neither FAP nor macro neighbour cells are detected then the same histogram method
of the reported FAP UE measurements will be used to limit the maximum downlink and
uplink transmit powers.
It is important to note that by setting the maximum uplink and downlink transmit powers it does not have
any effect on the power control loops and functionality that are active within the UMTS uplinks and
downlinks:

2.2.3 Admission Control
Admission control needs to check, before admitting a new UE, that the admittance will not sacrifice the
planned coverage area and the quality of the existing connection. Admission control accepts or rejects a
request to establish a radio access bearer in the radio access network. The admission control algorithm is
executed when a bearer is set up or modified. The admission control algorithm estimates separately in
uplink and downlink the load increase that the establishment of the bearer would cause in the access
network. The requesting bearer can be admitted only if both uplink and downlink admission control admit
it, otherwise it is rejected because of the excessive interference that it would produce in the network.
2.2.3.1 FAP capacity
In the context of the FAP system, Admission Control function is handled by the FAP RRC, and the
admission policy is dictated by its limited processing resources. The FAP can serve up to a maximum of 4
users depending on the mix of services of the users on a first come first serve basis. For example, the
FAP can hold up to 4 simultaneous Adaptive Multi Rate (AMR) calls, but if a 384/64 Packet Switched call
is active then only 2 new AMR calls can be admitted. However, admission control allows modification of
the service rate of existing PS calls to accommodate extra users. For example a 384/64 Packet Switched
connection will be modified if need to accommodate up to 3 AMR users.
In addition, for new Packet Switched type connections, FAP admission control attempts to verify that the
default (or requested ) downlink/uplink data rates can be sustained in the current radio environment of the
UE based on the UE reported CPICH Ec/No measurement reported on the RACH during call set up and
the FAP measured SIR during call setup.
In the local FAP case, a mixture of data and real time traffic Admission control must be established as
data traffic is seen from the radio side as a call and consumes bandwidth on the backhaul that might
affect the performance of other real time traffic. Admission Control mechanisms complement the
capabilities of QoS tools to protect real time traffic from the negative effects of other voice/data traffic and
to keep excess voice/data traffic off the backhaul network.
If Admission Control is enabled, new calls are allowed only if there are enough DPCH channels,
bandwidth and processing power left to be able to handle the resulting media streams effectively. This
Admission Control mechanism is a proactive mechanism that reduces the user possible bad service
perception at the beginning.
If the unit is being overloaded it is better to deny further calls, and thus maintain quality of the ones
currently in progress, rather than allowing too many calls that would overload the unit and reduce quality
of all calls in progress. This must be associated with calls priorities defining what call can be rejected or
redirected and with dynamical congestion control mechanisms.

NEC Confidential 5
th
J an. 2011
30

2.2.3.2 Summary of Capacity Management
As a summary, the following table list all methods and tools that FAPs use for Admission control handling:
Method Description
Admission Control Reject new User Requests
Redirect New User Request to another UMTS carrier or
neighbour GSM Cell
Switch the Channel type of existing PS call from DCH to
FACH (Release 2.0)
Reduce the data rate of existing PS call (Release 2.0)
Give priority to Emergency call
Table 3: Admission Control
2.2.3.3 New Call Handling at full capacity
When the FAP is running at its full capacity, e.g. 4 simultaneous voice calls at AMR 12.2 Kbps, it can use
following mechanisms to handle the 5
th
voice mobile originated call request:
- Redirection at Radio Level: Redirect the Radio Resource Control (RRC) connection request of CS
MO call to 3G/2G macrocell
- Channel Type Switching to Cell FACH (Release 2.0): Push the existing PS call in Cell DCH state to
Cell FACH state in order to accommodate a new CS call request
- Reduce the data rate of existing PS call (Release 2.0): Reconfigure the data rate of existing PS call
to lower data rate
- Directed Retry (Release 3.0): Reject the Call Request from Gateway with cause directed retry and
Redirect the RAB Assignment establishment request and relocate to Macro Cell

Redirection at Radio Level
Call Establishment starts with the UE sending RRC Connection Request with appropriate cause value.
When the FAP receives the message, it checks if there is any room to accommodate this new call request.
In case that the FAP is running at its full capacity, it rejects the request by sending RRC CONNECTION
REJECT (with cause: Congestion) and with necessary parameters to redirect the call request to either 3G
Macro Cell or 2G Macro cell. In the case that the 5
th
RRC Connection Request is placed for CS MO call,
then the FAP will still redirect the request to Macro Cell. In case that 5
th
RRC Connection Request is for
CS MT call or PS MO/MT call, then the FAP will reject the request and UE will stay in idle state in the FAP.
Channel Type Switching to Cell FACH
When the FAP receives the message, it checks if there is any room to accommodate this new call request.
In case that new RRC Connection Request for CS MO call and the FAP is running at its full capacity with
one of the occupied DPCH resources used by PS RAB, FAP free up DPCH resource by Channel Type
Switching of PS RAB to Cell FACH state.
Reduce the data rate of existing PS call
When a new PS connection is requested, call admission control will check the current usage of downlink
(DL) OVSF codes and uplink (UL) spreading factor resources in the modem. If sufficient resources are
available, it will attempt to assign the new user a 384k DL / 64k UL bearer, unless the user is R5 capable
in which case they will be assigned HSDSCH DL / 128k UL. If insufficient resources are available to
assign these high data rates, then the user will be assigned the maximum possible data rate based on

NEC Confidential 5
th
J an. 2011
31

the remaining available resources. If no resources are available for the minimum PS service (64k DL / 64k
UL), Call admission control will search the list of current users to find candidates that can be reconfigured
to lower data rates.
Directed Retry mechanism
In this mechanism, the FAP does not reject the UE at radio signalling level, i.e. by sending the RRC
CONNECTION REJECT, instead the FAP sets up the RRC connection with the UE. On receiving the
RAB ASSIGNMENT REQUEST message from Core Network, the FAP rejects the request with cause
directed Retry to perform the handover to macrocell. In this case the FAP initiates the Relocation
Preparation procedure as follows:
- The FAP terminates the request indicating unsuccessful RAB configuration assignment with the
cause "Directed retry".
- The RAN GW reports the outcome of the procedure in one RAB ASSIGNMENT RESPONSE
message.
- The FAP invokes relocation by sending RELOCATION REQUIRED message to the Core Network
with the cause "Directed Retry".
- The Core Network terminates the RAB Assignment procedure at reception of the RAB
ASSIGNMENT RESPONSE message.
- The Core Network establishes the relocation procedure with target macrocell.

Figure 1411 : Directed Retry

NEC Confidential 5
th
J an. 2011
32

2.2.4 Capacity and Multi-RAB Combination
The following table shows the services and RAB combinations supported by the FAP. The FAP will
support multi RAB allowing concurrent voice and data calls. The QoS mechanism in place will prioritise
voice connections over and above the data connections.

RAB Type
Max Users (all users on
same RAB combination)
AMR 12.2k 4
Video 64k 3
Packet Switched Downlink :64k/Uplink:64k 4
Packet Switched Downlink:128k/Uplink:64k 4
Packet Switched Downlink :384k/ Uplink :64k 2
Packet Switched HSDPA/64 4
AMR + Packet Switched 64/64 4
AMR + Packet Switched HSDPA/64 4
Video + Packet Switched 128/64 2
Video + Packet Switched 64/64 2
Video + Packet Switched HSDPA/64 2
Video + Packet Switched HSDPA/128 1
Table 7 : Services and RAB Combinations

NEC Confidential 5
th
J an. 2011
33

2.3 HSDPA
The NEC femtocell solution currently supports HSDPA 7.2 Mbps. HSDPA was introduced in UMTS R5;
the feature introduces the HSDCH which essentially is a downlink channel shared by multiple UEs UEs
are allocated downlink resources within the HS-DSCH in real time, depending on the presence of
downlink data for them.
2.3.1 Expected Throughput
For a HSDPA 7.2 Mbps bearer, there are a set of protocol and other overheads that will limit the rate that
a user experiences.

Therefore the absolute maximum theoretical data rate is 6.72 Mbps. However this does not take account
of:
- Hybrid Automatic Request Request (HARQ) re-transmissions.
- RLC status PDUs.
- RLC re-transmissions.
Even in very good channel conditions, there will always be a residual block error rate, so the HSDPA
scheduler can trade off power against block error rate to optimise overall throughput and minimise
interference; this means that in all practical scenarios, there will always be a certain block error rate that
reduces the observed throughput.
Therefore the maximum practical data rate for a HSDPA 7.2 Mbps carrier in good laboratory conditions,
when measured at the IP layer by a personal computer with a HSDPA data card, is approximately 6.4
Mbps.
The HSDPA 7.2 Mbps implementation supports up to 10 codes and 2 users per Transmission Time
Interval (TTI). Up to 4 active users are supported, multiplexed in time across multiple TTIs.
HSDPA 7.2 Mbps service begins to hit limitations on the processor power of the embedded processor.
Also, HSDPA service from a FAP is still subject to RF issues of interference and the changing channel
conditions of the UE, which can make the peak rate vary.
RLC header size: 3 bytes (AM PDU)
MAC header size: 4 bytes (for HSDPA)
The RLC/MAC total packet size is 656 bits
(this is the transport block size configured)
Therefore Payload size per MAC PDU is: 656 ((4 + 3)*8) = 600 bits
A 1500 byte TCP packet will be split into: 1500 / (600 / 8) = 20 RLC PDUs
FAP can send up to 112 MAC PDUs every 10ms to a category 8 UE
Maximum Payload transfer rate: (600 * 112) / 0.010 = 6.72 Mbps

NEC Confidential 5
th
J an. 2011
34

2.3.2 HSDPA Scheduler in FAP
Resources are allocated in terms of transport block size, modulation, channelization codes and transmit
power in particular TTIs and are signalled to the receiving UE by the femtocell. The entity in the femtocell
which selects these parameters is called the scheduler, and is implemented within the MAC-hs entity of
the PS-mode HSDPA protocol stack.
Regarding the tuning and control of the scheduling mechanism, it is very important to note that:
- Whilst in a macrocell, which will have large numbers of contending users, the tuning and efficiency of
the HSDPA scheduler is of paramount importance.
- In a femtocell, which has at most 4 HSDPA users and most likely will only have one or two in the vast
majority of cases, the tuning and efficiency of the HSDPA scheduler is far less important.
In all cases, each scheduled user is allocated transmission resource by the MAC-hs scheduler on a per-
TTI basis - this defines the number of codes to be used, the modulation scheme and the transmit power.
The selection of these resources is based on the UE's reported Channel Quality Indication (CQI) and
limited by their UE category (each category places a limit on the number of codes the UE can handle and
some categories can only support Quadrature Phase-Shift Keying (QPSK). It supports both QPSK and
16 Quadrature Amplitude Modulation (16QAM) and can transmit different modulation schemes
simultaneously to different users.
The MAC-hs implements a proportional fair metric based scheduler the metric for each user is derived
from their average throughput and their reported maximum instantaneous data rate. This means that:
- For multiple UEs the amount of resource allocated is according to their metric value (bigger metric
= more resource).
- If a UE is receiving downlink data, then their metric will go up and they will get more allocation.
Therefore more active users get more allocation and less active users get less allocation. If all users are
equally active their metrics will converge and their allocations will likewise converge.
Any pending re-transmission for a UE is automatically selected ahead of making a new initial transmission.
2.4 HSUPA
HSUPA is the next evolution step for 3G/UMTS networks. The HSUPA which is also known as FDD
Enhanced Uplink (EUL) has been introduced in the release 6 of 3GPP standards. Objective of HSUPA is
to enhance uplink packet data transmission by achieving data rates of up to 5.76 Mbps. Furthermore,
HSUPA will increase uplink capacity and reduce latency. A combination of HSDPA and HSUPA is
especially beneficial, since it will allow optimized packet data transfer in downlink and uplink. Services
that benefit from HSUPA are multimedia applications requiring excellent uplink performance, e.g. gaming,
video streaming, file upload. There are following UE categories on support of HSUPA. HSUPA would be
supported in Release 2.0 and would support Categories 2 & 3 at that release

HSUPA Category Max Uplink Speed
Category 1 0.73 Mbit/s

NEC Confidential 5
th
J an. 2011
35

HSUPA uses an uplink Enhanced Dedicated Channel (E-DCH) on which it employs link adaptation
methods similar to those employed by HSDPA including:
Higher-order modulation in addition to the existing Quadrature Phase Shift Keying (QPSK) - a
phase modulation algorithm - and 16-QAM (Quadruple Amplitude Modulation) (four amplitudes and
four phases) is used enabling higher data rates under favourable radio conditions (i.e. over less
noisy channels)
Shorter TTI (Transmission Time Interval) enabling faster link adaptation
HARQ (Hybrid ARQ (Automatic Repeat Request)) with incremental redundancy making
retransmissions more effective
Similarly to HSDPA, there is a packet scheduler, but its operate on a request-grant principle where the UE
(User Equipment) requests permission to send packets and the scheduler decides when and how many
UEs will be allowed to do so. A request for transmission will contain data about the state of the
transmission buffer, the queue at the UE and its available power margin.
2.4.3 HSUPA New Channel
Within the HSUPA framework, the Enhanced Dedicated Channel (E-DCH) is introduced as a new
transport channel for carrying user data on the uplink. On physical layer level, this translates into 2 new
uplink channels:
E-DCH Dedicated Physical Data Channel (E-DPDCH), and
E-DCH Dedicated Physical Control Channel (E-DPCCH).
The E-DPCCH carries control information associated to the E-DPDCH. In the downlink, 3 new channels
are introduced for control purposes:
- E-AGCH: E-DCH Absolute Grant Channel carrying absolute grants
- E-RGCH: E-DCH Relative Grant Channel carrying relative grants
- E-HICH: E-DCH Hybrid ARQ Indicator Channel carrying ACK/NACK
2.4.3 HSUPA Uplink Scheduling
The FAP maximum E-TFC selection mechanism takes into account the following for the HSUPA link
adaptation.
- scheduling information (UE power headroom, total E-DCH buffer status, Highest Logical Channel ID,
Highest Priority Logical Channel Buffer Status)
- QoS (User Priority and Guaranteed Bit Rate)
- reference power offset (E-DPDCH power associated to the reference E-TFC/DPCCH)
- the happy bit of E-DPCCH
- the uplink current cell load

The uplink scheduling mechanism is of central importance for HSUPA. As with HSDPA, the FAP supports
a proportional fair metric-based scheduling algorithm Task of the uplink scheduler is to control the uplink
resources the UEs in the cell are using. The scheduler therefore grants maximum allowed transmit power
ratios to each UE. This effectively limits the transport block size the UE can select and thus the uplink
data rate. The scheduling mechanism is based on absolute and relative grants. The absolute grants are
used to initialize the scheduling process and provide absolute transmit power ratios to the UE, whereas
the relative grants are used for incremental up- or downgrades of the allowed transmit power.

NEC Confidential 5
th
J an. 2011
36

The max E-TFC selection mechanism takes into account following:
- The scheduling information (UE power headroom, total E-DCH buffer status, Highest Logical
Channel ID, Highest Priority Logical Channel Buffer Status.
- The QoS (User priority and GBR)
- The reference power offset (E-DPDCH power associated to the reference E-TFC/DPCCH).
- The happy bit of E-DPCCH.
- The UL current cell load.

2.4.4 HSUPA HARQ
The HARQ protocol is a retransmission protocol improving robustness against link adaptation errors. The
FAP can request retransmissions of erroneously received data packets and will send for each packet
either an acknowledgement (ACK) or a negative acknowledgement (NACK) to the UE. The HSUPA
release supports incremental redundancy and chase combining HARQ types. There will be 4 HARQ
processes when TTI=10ms and 8 HARQ processes when TTI=2ms.
2.4.5 HSUPA Power Management
A mechanism is implemented in the FAP to prevent from downlink power shortage when HSDPA is
configured for the downlink HSUPA channels E-AGCH, Enhance Relative Grant Channel (E-RGCH) and
E-HICH. A fixed transmission power is allocated to these downlink HSUPA channels. Power is not
dynamically allocated between HSDPA and HSUPA downlink channels.
E-DPDCH power offset for HARQ power are configurable by the operator over the Layer 1 API. The
Node B Application Protocol (NBAP) optional parameters for E-AGCH, EHICH and E-RGCH power
offsets are not supported a fixed transmission power is allocated to these downlink HSUPA channels.

Femtocell Solution Radio Access Network Features

Uploaded by

Copyright:

Available Formats

Femtocell Solution Radio Access Network Features

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Femtocell Solution Radio Access Network Features

Uploaded by

Copyright:

Available Formats

NEC Femtocell Solution

Radio Access Network Features

You might also like