【资料名称】：爱立信HSPDA技术英文版

【资料作者】：ERI

【资料日期】：2011

【资料语言】：中文

【资料格式】：PPT

【资料目录和简介】：

WCDMA Evolved: HSDPA High speed downlink packet access
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
Broadband is clearly established as one of the fastest growing new services in history
End-user behaviour & Terminal availability
The Personal Computer goes mobile A mobile terminal is not always a mobile phone
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
HSDPA - Improving the WCDMA downlink
End user benefit: File Download Performance
Latency (radio-access network)
HS-DSCH coverage better than R99
No need for new sites
End user data rate is adapted based on radio conditions Downlink
High output power for the HS-DSCH channel
With link adaptation the lowest supported transport format is close to 64 kbps
Fast Hybrid ARQ with soft combining
Radio dependent scheduling (I.e. targeting fading peaks)

Uplink
Uplink 64 kbps which is same as for Release 99

HSDPA EVOLUTION
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
Release ’99 DL capabilities
DL packet traffic in R’99 of WCDMA
DCH transmitted on DPCH
Fixed SF (SF determines the channelisation code).
Variable data rate is achieved by: DTX (gating rate 1500 Hz).
Power controlled, support for SHO, highest rate – 2 Mbps.
DSCH transmitted on PDSCH
Variable SF.
Always DCH associated.
DSCH is shared by several users (single or multi-code transmission).
Power controlled (DPCCH), no support for SHO.
FACH transmitted on S-CCPCH
Fixed SF.
No power controlled (relatively high power), no support for SHO.
HSDPA key characteristics
HSDPA key characteristics
The physical layer retransmission (ARQ function)
R’99: RNC-based ARQ
HSDPA: Physical layer fast H-ARQ
Packet scheduling
R’99: RNC-based
HSDPA: NodeB-based
Power control
R’99: Fast PC, dynamic range: 20 dB in the DL, 70 dB in the UL
HSDPA: Link adoption function and AMC (adaptive modulation and coding)
AMC: to select coding and modulation scheme that requires higher Ec/Ior, which is available for the user close to the NodeB.
Good spectral efficiency.
HSDPA: up to 15 multi-codes in parallel.
Basic Principles
Shared Channel Transmission
New transport channel type, using multicode transmission
Radio resources dynamically shared among multiple users in time & code domain
Efficient code utilization

Higher Order Modulation
16QAM may be used as a complement to QPSK
16QAM allows for twice the peak data rate compared to QPSK
16QAM more sensitive to interference
=> Higher data rate in good radio channel conditions(high C/I, Little or no dispersion, Low speed)e.g. Close to cell site & Micro/Indoor cells
HSDPA modulation
16 QAM modulation is used (in R’99 – QPSK).
Phase and amplitude information is needed for proper channel estimation.
HSDPA capable terminal needs to obtain estimate of the relative amplitude ratio of HS-DSCH/CPICH power levels.
Channel coding
Only one transport channel is multiplexed on HS-DSCH (less redundancy).
Interleaving (TTI) at very short 2 ms rate.
Turbo coding (based on R’99) with a rate of 1/3.
Variation of the transport block size, the modulation scheme, and a number of multi-codes lead to other effective code rates (0.15 – 0.98).
HSDPA-capable NodeB is in charge of selecting the transport format to be used with particular modulation and number of codes.
Combination of modulation, coding scheme, and number of channelization codes => AMC (Adaptive Modulation and Coding).
Channel coding
Allowed combinations form TFRC (Transport Format and Resource Combination).
For intance: higher robustness is avaialble with a QPSK rate ¼ scheme, but at the penalty of only 119 kbps per code.
Given sufficiently good channel conditions, a single user may simultanously reveive 15 parallel multi-codes.
Short 2 ms Transmission Time Interval (TTI)
Reduced round trip delay on the air interface
Enables HSDPA features to operate at 500 times per second!
Fast Link Adaptation
Fast Radio Channel-dependent Scheduling
Fast hybrid ARQ with soft combining
Fast Hybrid ARQ with Soft Combining
H-ARQ principles (1)
Fast H-ARQ algorithm rapidly requests the retransmission of missing data entities, SAW (stop-and-wait) protocol is used.
Retransmitted data entities are soft combined with the original transmission before message decoding.
Since, the H-ARQ mechanism resides in the NodeB (MAC-hs), requests can be dome immediatly.
This way, probability of successful combining is increased.
If all data is correctly decoded, the ACK message is sent on the associated UL channel (HS-DPCCH).
H-ARQ requires some memory in the UE to buffer the soft information.
Two strategies of H-ARQ:
IR (incremental redundancy)
CC (chase combining)
H-ARQ principles (2)
The idea of the CC is to transmit an identical version of an errorneously detected data packet.
Then, receiver combines these two copies weighted by the SNR prior to decding, combining loss (0.2 – 0.3 dB per transmission).

The idea of the IR is to transmit additional redundant information that is incrementally transmitted if the decoding fails on the first attempt.
Causes increase of the effective coding gain with the number of retransmissions.
Full IR (inlcudes parity bits in every coded word) requires significant UE memory capabilities.
Chase Combining
Coding is applied to transmission packets
Soft combining of original and retransmitted signals is done at receiver before decoding
Advantage:
self decodable, time diversity, path diversity
Disadvantage:
wastage of bandwidth
Incremental Redundancy
Error Detection
Advantage:
Reducing the effective data throughput/bandwidth of a user and using this for another user
Disadvantage:
non-self decodable
Fast Link Adaptation (I)
Fast Link Adaptation (II) HS-DSCH Power Utilization
Fast Radio channel dependent Scheduling (I)
Scheduling = which UE to transmit to at a given time instant
Basic idea:
Transmit to users based on radio channel quality, targeting fading peaks
May lead to large variations in data rate between users
Tradeoff: fairness vs. cell throughput
Fast Channel-dependent Scheduling (II)
Examples of scheduling algorithms
Round Robin (RR)
Cyclically assign the channel to users without taking channel conditions into account
Simple but poor performance
Max C/I
Assign the channel to the user with the best channel quality
High system throughput but not fair
Proportional Fair (PF)
Assign the channel to the user with the best relative channel quality
High throughput, fair
Architectural Impact
Fast adaptation to changing radio conditions  new functionality in RBS!
HSDPA impact on RAN
Additional intelligence (HSDPA MAC – MAC hs) is installed at the NodeB.

Retransmission controlledby the NodeB leads to faster execution and shorter delay in case of retransmissions.

The Iub interface (NodeB-RNC) requires a flow control mechanism to ensure that NodeB buffers are used properly and there is no buffer overflow.

HSDPA impact on RAN
RNC still retains the RLC functionalities (providing retransmission in cases when HS-DSCH NodeB retransmission fails).

New MAC functionality at the NodeB (MAC-hs) handles the ARQ, scheduling, and priority handling.
HS-DSCH Protocol Architecture
Radio Interface Protocol Architecture
HSDPA key characteristics
HSDPA physical channels (1)
HS-DSCH (High Speed DL Shared Channel)
Carries the user data in the DL.
Higher modulation scheme (16QAM), lower encoding redundancy leading to high peak data rates.
TTI (Transmission Time Interval), interleaving period = 2 ms(In R’99, TTI = 10/20/40/80 ms).
Fixed SF (16), support multi-code transmission, as well as multiplexing of different users (15 – maximum capability, depends on the UE: 5/10/15).
Users check the information on the HS-SCCH to determine which HS-DSCH codes to despread.
HS-DSCH has always DL DPCH associated (signal radio bearer for layer 3 signaling, power control command for UL HS-DPCCH, etc.)

HSDPA physical channels (2)
HS-SCCH (High Shared Control Channel)
Carries the information needed for HS-DSCH demodulation.
The UTRAN allocates a number of HS-SCCHs corresponding to the maximum number of users code-multiplexed.
If there is no data on HS-DSCH, HS-SCCH is not assigned.
The HS-SCCH uses SF 128, accomodating 40 bits per slot.
Each HS-SCCH block has a three–slot duration divided into 2 functional parts:
First part (first slot) carries the time-crucial information needed to start the demudaltion process in due time ( -> avoid chip level buffering) and indication if QPSK or 16QAM modulation is used on HS-DSCH.
Second part (next two slots) contains CRC (cyclic redundancy check) for checking HS-SCCH, ARQ process number, redundancy version.
HSDPA physical channels (3)
HS-DPCCH (UL High Speed Dedicated Physical Control Channel)
Carries ACK/NACK information for the L1 retransmissions.
Carries CQI (DL Channel Quality Indicator) to be used by NodeB scheduler to determine to which terminal to transmit and at which rate.

Intensively discussed in the 3GPP forum, feedback method is not easy to be standardized due to differences in the terminals.
The feedback information consists of 5 bits.
One state expresses: ’do not bother to transmit’.
Other states represent the transmission that terminal can receive.
These states range from single code QPSK transmission to 15 mulit-code 16QAM (including various coding rates).
Terminals, which do not support certain number of codes should signal this for power-reduction factor related to the mose demanding supported TFRCs.
Physical layer operation procedure (1)
The scheduler in the NodeB evaluates channel conditions for each UE.

The NodeB identifies the HS-DSCH parameters (number of codes/modulation/H-ARQ mode).

The HS-SCCH is sent two slots before the corresponding HS-DSCH.

The UE monitors the HS-SCCH, and once the first part in decoded, the UE starts bufffering the necessary codes from the HS-DSCH.

When the second part of the H-SCCH is decoded, the UE determines to which ARQ process the data belongs and if needed – it is combined with the data in the soft buffer.
When data are decoded, the UE send in th UL an ACK/NACK indicator (depending on the CRC check conducted on the HS-DSCH).
If the NT continues to transmit data for the UE in consecutive TTIs, the UE will stay on the same HS-SCCH.
Physical layer operation procedure (2)
Terminal sends the ACK/NACK in the UL HS-DPCCH 7.5 slots after the end of the HS-DSCH TTI.
The network side is asynchronous in terms of when to initiate the DL transmission.
Depending on the implementation, different times might be spent on the scheduling process.
Physical layer operation procedure (3)
DL DCH, and UL DCH (R’99 channels) are not slot-aligned to the HSDPA transport channels. HS-DSCH and HS-DPCCH may start in the middle of the slot.
The UL timing is quantised to 256 chips (symbol-aligned)
Mobility with HSDPA
UTRAN determines the serving cell for the HSDPA capable UE.
Handover is synchronized, so the HS-DSCH connectivity is maintained.
Stop and start of the transmission is done at a certain time dictated by the UTRAN.
The serving HS-DSCH cell may be changed without updating active set for the R’99 channels.
Thus – new measurement event is included in R’5 informing the UTRAN of the best serving HS-DSCH cell.
Intra-Node B Handover (between sector)
Hard handover in the HS-DSCH, with most likely softer handover of the DPCH and the UL HS-DPCCH.
Serving HS-DSCH cell change
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
Why Ericsson?
Ericsson 3G offering builds on true 3G platforms & products already from the start

Ericsson solutions provide cost efficient upgrade to HSDPA
RNC: Only SW upgrade needed
RBS 3000 family:
SW & limited BB HW upgrade needed (existing customers)
Integrated & cost efficient transmission solutions

HSDPA support in all RBS 3000 products
RBS 3000 has Superior Base band Architecture
Optimal performance: All vital HSDPA functionality on same board
Scheduling, Power allocation, Link-adaptation, Flow control
Optimal dimensioning:
HS-TXB & RAXB supports both DCH & HSDPA traffic
DCH & HSDPA share the same hardware resources
Separate hardware for DL and UL
HSDPA channel structure & RAN architecture
Ericsson’s architecture even more beneficial in HSDPA
RBS 3000 family provides differentiation
RBS 3000 can be equipped with power
classes from 20 to 60W

The high power PAs give more available power for HSDPA

- Better end user experience through
higher peak rates

- Increased cell capacity

- Better HSDPA coverage => Higher bit rates on cell border & indoor

MCPA/TRX are HSDPA prepared
MCPAs can handle both QPSK and 16QAM modulation schemes at the same declared output power as for todays system






Cost efficient transmission solutions with Ericsson’s 3G platform, CPP
Best effort for HS in seperate AAL2 path
HSDPA VC on CBR optimizes load with best effort QoS and a flow control RBS-RNC to minimize re-transmission
UBR+ as an efficient alternative to CBR
A minimum rate is guaranteed... ... it may be exceeded whenever there is unused capacity... ... from any other connection within the same AAL2 Path (CBR VP)
Efficient use of dimensioned transmission capacity
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
HSDPA Phasing
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
Terminals
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
HSDPA Trial system
HSDPA live over the air
Integrated with commercial products
Up to 4.9 Mbit/s (QPSK & 14 codes)
Simultaneous speech and HSDPA
Main purpose
Allow for earliest possible commercial launch
Evaluation of radio characteristics (mixed R99 and R5)
Verify the lower layers of the standard (HW impact)
Knowledge build-up
Evaluate applications
Route Industrial Area
HSDPA – WCDMA Evolved
Mobile Broadband opportunities
What are the benefits with HSDPA?
How does HSDPA work?
Why Ericsson?
Roadmap
Terminals
Trial system
Summary
Summary – WCDMA Evolved High Speed Downlink Packet Access
Summary – WCDMA Evolved High Speed Downlink Packet Access
Thank you for your attention!

HSDPA performance (simulations [1] - 1)
Fully dynamic simulations, code resource (HS-DSCH) sharing among few users is considered.
Base station have fixed power allocated to HS-DSCH (14 W), 3 sectored antennas equally spaced (2 km).
Two multiplexing cases are analyzed:
Semi-static: maximum number of multi-codes that each HS-DSCH user can have is fixed (n). E.g., for 15 multi-codes and 5 codes-multiplexed users in the cell, each can have maximum 3 multi-codes (depending on the channel conditions).
Dynamic code-sharing: the user in a cell with n code multiplexed users can have maximum available multi-codes. E.g, for 15 multi-codes, 8 codes can be allocated for one user (due to channel conditions), then second user in the queue would have maximum 7 codes avaialble and so on.
HSDPA performance (simulations [1] - 2)
HSDPA performance (simulations [1] - 3)
HSDPA / R’99 DSCH ( simualtions [2] – 1 )
Dynamic system level simulations, hexagonal layout consisting of 19 sites equally spaced (2700 m).
Path loss: L = 128.1+37.6log(R), shadowing std: 8 dB.
Fast fading: ITU Pedestrian A channel model at 3 km/h.
Traffic: packet calls (mean size of 25 kbps).

R’99 DSCH, multiple users are code-multiplexed, the time multiplexing is done at the frame level (10 ms TTI).

R’5 HSDPA, maximum 12 multicodes can be allocated to the single UE (8,64 Mbps – peak data rate), H-ARQ scheme with CC (chase combaning) is used. DL radio conditions are evaluated by the CPICH Ec/N0 measurements made by the UE.
HSDPA / R’99 DSCH ( simualtions [2] – 2 )
Total amount of data successfully transmitted by the sector during simulation duration.
HSDPA performance (simulations [3] - 1)
Short-term dynamic simulation concept.
Introduces a short period of dynamic simulation after each snapshot.
Within the short period dynamic relevant aspects for radio network planning are taken into account.
Time dependent fading effects (fast fading) are not taken into account, RRM mechanism don’t make decisions based on short-time changes in the propagation environment.
Traffic model includes packet http users (TCP flow control neglected – results on call level only).
Different scheduling methods are considered:
Equal time: serves all users requesting data in a round robin manner one after other for one TTI.
Equal data rate: weighted round robin where the weights are applied according to the inverse of the estimated throughoput.
Maximum throughput: prioritizes users with the best channel conditions.
HSDPA performance (simulations [3] - 2 )
HSDPA/DCH power allocation ( [4] - 1)
Dynamic simulations, layout: equally spaced (2,8 km), horizontally wide (700) antennas deployed in 3-sectored manner.
Each HSDPA user is assumed to receive the HS-DSCH, HS-SCCH, and one associated DPCH (only RRC signalling @ 3 kbps).
Cost-231-Hata propagation model, ITU Vehicular Achannel.
Traffic layer: 64 kbps DCH users, bit rate for HSDPA is controlled (MAC-hs packet scheduler).
Offered traffic is adjusted so that there is full utilization of the HS-DSCH in every cell, and thus all the available power for DCH transmission is used.
HSDPA/DCH power allocation ( [4] - 2)
CDF of the total transmit power per cell depending on the allocated HSDPA power
HSDPA/DCH power allocation ( [4] - 3)

Enhancing WCDMA for TCP: HSDPA and Beyond
Recap from last year
“It’s about the latency, stupid” (see http://rescomp.stanford.edu/~cheshire/rants/Latency.html)
High Speed Downlink Packet Access improves TCP performance by
Retransmissions from NodeB
Smaller TTI
HARQ
Increased peak bit rate
Uplink not improved
TCP performance was estimated to double

Outline for today
Simulation results for HSDPA
available data rates
TCP performance
comparison to DCH
HSDPA lessons
Uplink improvements


Available Peak Data Rates
Average available data rates in a single cell system
Single ray Rayleigh: 6.9 Mbps
Pedestrian A: 3.7 Mbps
Vehicular A: 1.6 Mbps
In a real multi-cell system slightly reduced rates, e.g. for Ped A 3 Mbps
TCP Performance: Object transfer times
Comparison with DCH
HSDPA is Good!
Simulation results show that HSDPA improves performance
significantly higher peak data rates available
RTT reduced
Performance improvement
Transfer delay halved
Per packet bit rate doubled
As expected last year
Delays limit the performance

Reality Check
Expect a reduced performance
TTI increased from 0.67 ms to 2 ms
HARQ round trip time (roughly) 10 ms
HARQ signaling errors result in RLC retransmissions
Higher modulation schemes optional
Processing times tend to increase as product launch approaches...
In a loaded system, enhanced performance can be traded for increased capacity
Bottom line: HSDPA allows operators to provide significantly improved packet data access if they wish so

TCP Delays
Uplink contributes significantly to the total round trip time
60-70% with no Internet delay
25-30% with 50 ms Internet delay
Improving uplink
Various methods being discussed
HARQ
Retransmissions from NodeB
Shorter TTI
Coverage increased for current data rate (theoretically up to 4 Mbps)
The effect of these methods not known (yet), so let’s assume that the improvement of the uplink reduces delay by 85%
HSUPA Performance: Comparison with DCH
HSUPA Conclusions
Details pretty much open
Potential gain of 20%-40% in downlink performance
Uplink performance also important
Video phones
Pictures, multimedia from terminals
Sending email
Interactive gaming
First studies ready by the end of the year, more studies next year
HSDPA/HSUPA Packet Scheduling
JARNO NIEMELÄ
jarno.niemela@tut.fi

21.03.2005
Outline
Principles of packet scheduling in WCDMA / HSDPA Rel’05
Performance analysis of HSDPA PS for NRT services [1]
Scheduling in E-DCH/HSUPA (NRT services) [2]






Packet scheduling in WCDMA/HSDPA Rel’05
NodeB controlled packet scheduling (fast).
Sensitivity of throuhgput for channel quality
Task of packet scheduler
To schedule interactive and background services (NRT) for users.
To allocate radio resources efficienctly for a cell such that cell capacity will be maximized while fulfilling the QoS requirements according to certain policy.
To monitor allocation of NRT services and system loading.
To perform load control actions.
Input parameters for packet scheduler
Resource allocation
HS-PDSCH and HS-SCCH powers
HS-PDSCH codes
Number of HS-SCCHs
Downlink channel quality measurements
CQI reports
Power measurements on associated DPCH
HARQ acknowledgements
QoS parameters
QoS attributes
Scheduling priority indicator (SPI)
Guarantee bit rate
Miscellaneous
Amount of buffered data
Mobile capabilities
Fairness
Selection of scheduling approach is always a trade-off between the fairness and maximum cell throughput.
C/I scheduling maximizes the system capacity with the cost of lack of fairness.
Fair resources scheduling distributes equally the radio resources (codes, power and allocation time). Not completely fair.
Fair throughput tries to provide the same throughput for all users.

Packet scheduling algorithms
Slow scheduling methods (Blind)
Average C/I
Round robin
Fair throughput
Does not consider instantaneous radio conditions

Fast scheduling methods (Advanced/opportunistic)
Maximum C/I
Proportional fair
Fast fair throughput
Utilizes temporary changes of radio conditions
Slow scheduling methods
Average C/I (Avg. C/I)
Priorities users with the highest average C/I (~100 ms period)
Fast fading averaged out
Round Robin (RR)
Cyclic order used without considering channel conditions
Blind method
Simple and allocates radio resources evenly between the users (=high fairness)
Fair Throughput (FTH)
No instantaneous channel information utilized
Priorities users with lowest average throughput
Fast scheduling methods (1/2)
Maximum C/I (Max. C/I)
Serves in every TTI (transmission time interval) the user with the best radio conditions with the largest supportable bit rate.
High cell throughput, low fairness.
Proportional fair (PF)
Serves the user with largest relative channel quality:



where Pi(t) denotes the user priority.
User’s with relatively good channel conditions are served. Available information of CQI and previous transmissions is utilized.
Fast scheduling methods (2/2)
Fast fair throughput (FFTH)
Aims at providing a fair throughput distribution among all the users in the cell, while still taking advantage of the fast fading variations



whereis the average supportable data rate of a user i and is a constant that indicates the maximum average supportable data from all j users.


Summary
Multi-user diversity
Fast allocation (2ms TTI) of radio resources  Users with good radio conditions served  Multi-user diversity (selection diversity)
Increases the system/
cell throughput
The gain naturally
depends on the dynamics
of fast fading (short term
variations)

Throughput vs. Es/N0
Gain of multi-user diversity
Performance analysis of PS in HSDPA
User throughput distribution
Average user throughput
Efficiency of resource utilization
Fast scheduling is able to use more efficiently higher MCSs.
Link utilization
Performance of Max C/I and PF under high load
Cell throughputs (1/2)
With minimum user throughput guarantees (< 64 kbps)
Cell throughputs (2/2)
Summary table of cell throughputs with minimum user throughput guarantees
Conclusions from PS methods for HSDPA
Selection of PS algorithm important for HSDPA capacity maximization and QoS provisioning.
Multi-user diversity gain for 10-15 users 100 % in PedA and 50 % in VehA channels (over RR).
Max C/I maximizes the cell throughput (with degraded QoS provisioning)
Proportional fair scheduler seems to provide a trade-off between QoS and cell throughput (time dispersion of the channel still a great problem.
Fast packet scheduling for E-DCH/HSUPA
UL PS in REL’99
RNC –based packet scheduling
Upgrading based on capacity requests
Downgrading based on inactivity timer
PS approaches for Node B scheduling (1/2)
Blind data rate detection (BRD)
Instantaneous (TTI=10ms) data rate observed by Node B and compared to maximum allowed. This information is thereafter used for resource allocation according to UE´s actual needs.
PS algorithm based on resource utilization factor (RUF)


PS approaches for Node B scheduling (2/2)
Time Division Multiplexing (TDM)
Fast allocation (TTI=2ms) based on same approach as in HSDPA.
Easier to keep resource utilization closer to the planned one.
Exploitation of instantaneous channel conditions.
Requires uplink syncronization
Utilization of USTS (uplink synchronous transmittion scheme) [5]
Synchronization achieved through DL frames. Would require guard intervals together with using the information provided by RTT.
To support SHO, only one Node B is allowed to perform scheduling decisions.
Allocation strategies (RRFT, maximized transmit power efficincy (MTPE), PFT)
Performance analysis (macrocellular)
Performance analysis (macrocellular)
Performance analysis (microcellular)
Performance analysis (microcellular)
Performance analysis
Conclusions from PS for E-DCH/HSUPA
Node B PS based on BRD is able to provide 30-40% capacity gain over RNC based PS (TVM)
Intuitively, channel-dependent methods are able to provide better performance
Uplink synchronisation provides capacity gain of 20%.
Extra signalling load might reduce the capacity gains in some extent.

Main references
1. Pablo José Ameigeiras Gutiérrez, “Packet Scheduling and Quality ofService in HSDPA”, Ph. D. Thesis,AalborgUniversity, Denmark, October 2003.
2. José Outes Carnero, “Uplink capacity enhancements in WCDMA,” Ph. D. Thesis, Aalborg University, Denmark, March 2004.
3. H. Holma, A. Toskala (ed.), “WCDMA for UMTS,” 3rd ed., John Wiley & Sons, Ltd., 2004.
Accessory references
4. J. Laiho, A. Wacker, T. Novosad, “Radio Network Planning and Optimisation for UMTS,” John Wiley & Sons, Ltd., 2002.
5. 3GPP, “Study report of Uplink Synchronous Transmission Scheme (USTS),” TR 25.854, Ver 5.00, Rel. 5., December 2001.