This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
• HSDPA Improvements: – 64QAM (RAN1643)– MIMO (RAN1642)– MIMO 42Mbps with 64QAM (RAN1912)– Dual-Cell HSDPA (RAN1906)– DC-HSDPA with MIMO 84Mbps & 64QAM (RAN1907)– Flexible RLC in DL (RAN1638)– Dual Band HSDPA (RAN2179)
• HSUPA Improvements: – Frequency Domain Equalizer (RAN1702)– HSUPA Interference Cancellation Receiver (RAN1308)– HSUPA 16QAM (RAN1645)– Flexible RLC in UL (RAN1910)– HSUPA Downlink Physical Channel Power Control (RAN971)– Dynamic HSUPA BLER (RAN2302)
• HSDPA Improvements: – 64QAM (RAN1643)– MIMO– MIMO 42Mbps with 64QAM– Dual-Cell HSDPA– DC-HSDPA with MIMO 84Mbps & 64QAM– Flexible RLC in DL– Dual Band HSDPA
• UE peak data rate increased to 21.1 Mbps (L1 - theoretical)• Max application level throughput ~17.9 Mbps (ideal channel)• 64QAM is applicable for better radio conditions
MaxBitRateNRTMACDFlow can be used to restrict the maximum bit rate of NRT MAC-d flow. The bit rate used in the reservation of the resources for the MAC-d flow is the minimum value of 1) max. bit rate based on UE capability, 2) max. bit rate of the RAB, 3) activated HSDPA bit rate features and 4) the value of this parameter. This parameter does not limit the maximum instantaneous bit rate on air interface. The value of the parameter is compared to the user bitrate of the NRT MAC-d flow excluding MAC-hs header, RLC header and padding.WCELRNHSPA; 128..2112083968; 128; 65535*
*65535 parameter does not restrict the maximum bit rate,
but the maximum bit rate is restricted by other limits.Range & Default value changed with RU30 to: 0128...83968 ; 128; 0 =
same meaning as 65535 in RU20 (HSDPA peak rate not
• HSDPA Improvements: – 64QAM– MIMO (RAN1642)– MIMO 42Mbps with 64QAM – Dual-Cell HSDPA– DC-HSDPA with MIMO 84Mbps & 64QAM– Flexible RLC in DL– Dual Band HSDPA
• MIMO: Multiple-Input Multiple Output• M transmit antennas, N receive antennas form MxN MIMO system• huge data stream (input) distributed toward m spatial distributed antennas; m parallel bit streams
(Input 1..m)• Spatial Multiplexing generate parallel “virtual data pipes”• using Multipath effects instead of mitigating them
MIMOEnabledWCEL; 0 (Disabled), 1 (Enabled)• RU20 (3GPP Rel. 7) introduces 2x2 MIMO with 2-Tx/2-Rx
– Double Transmit on BTS side (D-TxAA), 2 receive antennas on UE side– System can operate in dual stream (2x2 MIMO) or single stream (Tx diversity) mode
• MIMO 2x2 enables 28 Mbps peak data rate in HSDPA – 28 Mbps peak rate in combination with 16QAM – 64QAM: no simultaneous support of 64QAM & MIMO (not yet)– Dual-Cell HSDPA: not possible to enable MIMO & DC-HSDPA in a cell in parallel
• Benefits: MIMO increases single user peak data rate, overall cell capacity, average cell throughput & coverage • UE categories for MIMO support: Cat. 15, 16, 17 & 18
Prerequisites: • double Power Amplifier units & antenna lines per cell; • must be enabled: HSDPAEnabled, HSUPAEnabled, HSDPA14MbpsPerUser, HSDPADynamicResourceAllocation,
FDPCHEnabled, HSDPAMobility, FDPCHEnabled, FRLCEnabled; must not be enabled: DCellHSDPAEnabled
• MIMO enabled cell: S-CPICH is broadcast for DL channel estimation in UE– S-CPICH transmission power is controlled with existing parameter
• UE must be able to estimate each of the 2 signals separately– P-CPICH is broadcast along with data stream 1– S-CPICH (new with RU20) is broadcast along with data stream 2– SF 256 spreading code must be allocated in DL to support S-CPICH transmission
MIMOS-CPICH Power & Code allocation
SF 16
SF 32
SF 64
SF 128
SF 256
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
P-CPICH
P-CCPCHAICH
PICH
S-CCPCH
depending on
FACH / PCH
configuration
HS-SCCH
E-RGCH
E-HICH
E-AGCH 10ms
E-AGCH 2ms
S-CPICH MIM
O
F-DPCH
,0
S-CPICH tx power =PtxPrimaryCPICH-10..50; 0.1; 33 dBm
• HSDPA Improvements: – 64QAM– MIMO– MIMO 42Mbps with 64QAM (RAN1912)– Dual-Cell HSDPA– DC-HSDPA with MIMO 84Mbps & 64QAM– Flexible RLC in DL– Dual Band HSDPA
Basics:• optional Feature; RNC License Key required (ON-OFF)• RU20 enables either 2x2 MIMO (RAN1642) or 64QAM (RAN1643)• RU30 enables simultaneous 2x2 MIMO and 64QAM operation (RAN1912)• Peak Rates: up to 2 x 21 Mbps = 42 Mbps• 3GPP Rel. 8• new UE Categories: 19, 20
Requirements• Flexible RLC, F-DPCH, MIMO 28 Mbps, HSDPA 64QAM
• Switching can occur when conditions change, i.e. when it becomes possible to support MIMO with 64QAM, or when it is no longer possible to support MIMO with 64QAM
• The conditions required to support MIMO 42Mbps with 64QAM are:– it must be possible to support MIMO– it must be possible to support HSDPA 64QAM– The WCEL MIMOWith64QAMUsage parameter must be set to enabled– The BTS and UE must support simultaneous use of MIMO and 64QAM
• If MIMO with 64QAM is not possible but MIMO without 64QAM, or 64QAM without MIMO is possible, MIMO shall be preferred
• HSDPA Improvements: – 64QAM– MIMO– MIMO 42Mbps with 64QAM – Dual-Cell HSDPA (RAN1906)– DC-HSDPA with MIMO 84Mbps & 64QAM– Flexible RLC in DL– Dual Band HSDPA
• DC-HSDPA provides greater flexibility to the HSDPA Scheduler, i.e. the scheduler can allocated resources in the frequency domain as well as in the code and time domains
RNC license for specific number of cells• following features must be enabled:
• HSDPA (HSDPAEnabled)• HSUPA (HSUPAEnabled)*• HSDPA 15 codes (HS-PDSCHcodeset) • HSDPA 14 Mbps per User (HSDPA14MbpsPerUser)• HSDPA Serving Cell Change (HSDPAMobility)• Fractional DPCH (FDPCHEnabled)• DL Flexible RLC (FRLCEnabled)• Shared Scheduler for Baseband Efficiency• HSPAQoSEnabled must be configured with the same value in both DC-HSDPA cells• MaxBitRateNRTMACDFlow (def. 65535 = not restricted) should be configured to allow the peak throughput• RU20: MIMO + DC-HSPDA must not be enabled for all cells belonging to the Node B
(MIMOEnabled); ; • RU340: MIMO + DC-HSDPA possible DC-HSDPA + MIMO possible in RU340
• A single HSDPA shared scheduler for baseband efficiency is required per DC-HSDPA cell pair• 3 HSDPA shared schedulers are required for a 2+2+2 Node B configuration with DC-HSDPA• Each scheduler is able to serve both HSDPA & DC-HSDPA UE on both RF carriers• Link Adaptation is completed in parallel for each RF carrier
• HSDPA Improvements: – 64QAM– MIMO– MIMO 42Mbps with 64QAM – Dual-Cell HSDPA– DC-HSDPA with MIMO 84Mbps & 64QAM (RAN1907)– Flexible RLC in DL– Dual Band HSDPA
Benefits:• higher Peak Rate: up to 2 x 2 x 21 Mbps = 84 Mbps• better Coverage due to DC-HSDPA & MIMO• More robust transmission due to MIMO & DC HSDPA usage
Basics:• enables simultaneously: DC HSDPA, MIMO & 64QAM• MIMO uses Single Stream or Double Stream
transmission• DC-HSDPA uses 2 cells (in 1 sector) at same BTS; same frequency band & adjacent carriers to a UE• 64QAM 6 bits/symbol
Requirements• RAN1642 MIMO 28 Mbps• RAN1638 Flexible RLC • RAN1906 DC HSDPA• RAN1643 64QAM• RAN1912 MIMO 42Mbps +
64QAM
DC-HSDPA with MIMO & 64QAM
DC-HSDPA with MIMO (w/o 64QAM)
DC-HSDPA with 64QAM (w/o MIMO)
DC-HSDPA (w/o MIMO, 64QAM)
64QAM with MIMO (w/o DC-HSDPA)
UE Categories(3GPP Rel. 9; TS 25.306)
MaxBitRateNRTMACDFlow* can be used to restrict max. bit rate of NRT MAC-d flow WCELRNHSPA; 0128... 83968 ; 128; 0value 0 / 65535 (before): HSDPA peak rate not limited by the RNC
• DC-HSDPA with MIMO can be maintained, activated or de-activated during mobility• Availability of DC-HSDPA with MIMO checked in target cell when SCC or HHO initiated• If DC-HSDPA with MIMO cannot be used in the target cell mobility proceeds without it:
– DC-HSDPA or MIMO is used if possible, according to the parameter DCellVsMIMOPreference • If HSUPA IFHO can be used DC-HSDPA & MIMO is not be deactivated but is maintained
during Inter-Frequency measurements• If HSUPA IFHO cannot be used, E-DCH to DCH switch is completed before inter-frequency
measurements; DC-HSDPA with MIMO is deactivated at the same time
• DC-HSDPA with MIMO is not supported across the Iur• S-RNC does not configure DC-HSDPA with MIMO if there are radio links over the Iur in the
active set
SCC: Serving Cell Change
DCellVsMIMOPreferenceWCELRNHSPA; DC-HSDPA preferred (0), MIMO preferred (1)
defines whether RNC primarily activates DC-HSDPA or MIMO for a UE, which supports both DC-HSDPA & MIMO in case simultaneous usage of DC-HSDPA &
• HSDPA Improvements: – 64QAM– MIMO– MIMO 42Mbps with 64QAM – Dual-Cell HSDPA– DC-HSDPA with MIMO 84Mbps & 64QAM– Flexible RLC in DL (RAN1638)– Dual Band HSDPA
• Prior to Rel. 7: RLC layer segments high layer data units (IP packets) in RLC PDU sizes of 336 and 656
– 336 is 320 net bit plus 16 bit RLC OH– 656 is 640 net bit plus 16 bit RLC OH
• On MAC-d layer did not increase Overhead– Data was passed directly to MAC-hs layer (MAC-d)
• Several MAC-d PDUs were concatenated to form a MAC-hs data block
• BTS selects proper MAC-hs data block size based on– available user date in BTS buffer and– radio conditions for that UE
• With DL Flexible RLC the RNC adapts the RLC-PDU size to the actual size of the higher layer data unit (IP)– maximum size of 1500 Byte is supported (IP packet length in Ethernet)
• Major improvements with DL Flexible RLC– less processing in RNC & UE– higher end user application throughput– lower latency for packet access– Significantly lower Overhead– Much less padding bits – Lower risk for RLC stalling because of too small transmission windows
Dual Band HSDPA: RAN2179• Included in RU40 application software package – license required• HW prerequisites: Flexi rel.2• Can be used if: DC-HSDPA and HSPA Peak Rate Upgrade features are active
Brief Description: • This feature introduces for a single UE the possibility of using simultaneously two carriers
in DL that are situated on two different WCDMA frequency bands• Feature enables achieving 42 Mbps peak rate for user in DL (assuming 64QAM and 15
codes usage on both frequencies)• Comparing to single carrier case gives possibility to increase cell throughput• Feature is much similar to DC-HSDPA in function• Feature restricts single carrier usage in UL (DB or DC-HSUPA is not allowed)
Motivations and Benefits:• High Throughputs – This feature enables throughputs as high as 42 Mbps.• Better Coverage – Dual Band allows using two different frequency bands. For cases
where high coverage is needed, lower Band of the two can be used to enhance coverage.• Configurations flexibility – This feature with carriers from 2 different frequency bands
Basics:• before RU30: Node B receiver based on RAKE receiver technology RAKE unable to receive high data rates even in total absence of other cell interference short spreading codes (SF2) vulnerable to ISI
• RU30 introduces:- RAN1702: Frequency Domain Equalizer FDE- RAN1308: HSUPA Interference Cancellation IC• FDE can remove ISI, leaving other users of same cell & surrounding
cells to be main limiting factors for UL data rates• Interference from other users of the own cell can reduced by HSUPA IC• FDE is prerequisite for UL 16QAM (RU430)
Frequency Domain EqualizerOnly Rrake receiver was used in RU20 & earlier releases
RAKE delivers adequate performance for data rates below 2 Mbps; its main tasks are: Identify the time delay positions at which significant energy arrives and allocate
correlation receivers, i.e. Rake fingers, to those peaks. Within each correlation receiver, track the fast-changing phase and amplitude
values originating from the fast fading process and utilize them. Combine the demodulated and phase-adjusted symbols across all active fingers
and present them to the decoder for further processing.
Frequency Domain EqualizerFDE = linear equalizer + fast convolution
• FDE (LMMSE) provides optimal linear estimate of transmitted signal accounting for both:• Channel impact (fading)• Interference + noise
• FDE is a combination of linear equalization & fast convolution. • Convolution is relatively demanding in terms of computation • Convolution can be replaced by multiplication if completed in the frequency domain FFT
• FDE reduces the effects of ISI arising from user’s own signal due to multipath propagation.• FDE applied to users with granted 2xSF2 + 2xSF4 (QPSK or 16-QAM) up to 11.5 Mbps.
FDE scheme
signal FFT
pilotChannel
estimation
MMSE filter coefficient calculation
IFFTDespreading
and detection
bits
Time domain
Frequency domain
(I)FFT: (Inverse) Fast Fourier TransformationISI: Inter-Symbol-InterferenceLMMSE: Linear Minimum-Mean-Square-Error
FDE sensitive to channel estimation => E-DPCCH boosted mode used for channel & SIR estimation E-DPCCH boosted mode E-DPCCH bound to E-DPDCH power (not to DPCCH, as usual) Starting from ETFCIBoost E-TFCI Default value, UE selects 16QAM & start to use boosted mode. Boosted mode introduction to increase E-DPCCH power proportionally to high data rates. High data rates are
source of high self interferences boosted E-DPCCH useful for E-DPDCH channel estimation & demodulation. not mandatory for UE to support E-DPCCH power boosting (requires Rel. 7 or newer UE) UE indicates support of E-DPCCH power boosting within RRC Connection Setup Complete message RNC signals E-DPCCH power boosting parameters to UE
Basics:• reduces UL Intra-cell interference with non-linear Interference Cancellation IC
method called Parallel Interference Cancellation (PIC)
• RAN1308: Basic PIC decreases interference from HSUPA 2 ms TTI users to other UL channels • improved coverage e.g. for AMR calls existing in parallel with peak rate users
• RAN2250: Enhanced PIC (RU450) decreases interference from HSUPA 2 ms TTI users on each other • larger peak HSUPA data rates (also 16-QAM)
2ms HSUPA Interference cancelled Non-IC users signal
(Residual signal)
“IC users”
“Non-IC users”
• UL signal received with Rake Receiver or FDE technique
• Turbo decoding obtain 2 ms TTI E-DCH signals
• Decoded data used to reconstruct original 2 ms TTI signals (interference for other users) Reconstruction includes turbo encoding spreading & modulation.
• Cancel interference from 2 ms TTI user: Reconstructed signals are summed up & subtracted from the original antenna signal non-IC users’ signal (residual signal)
• Non-IC users signals are demodulated on the residual signal, benefiting from a lower interference level improving cell coverage & capacity
FDE: Frequency Domain EqualizerPIC: Parallel Interference Cancellation
Enhanced PIC methodBasic PIC: IC users do not benefit directly from reduced interference
their signals are demodulated in parallel on the original antenna signal
Enhanced PIC (RAN2250): • demodulate IC users’ signals again after residual signal reconstruction for these
signals (to gain from IC of Basic PIC).
• Residual Stream Reconstruction RSR: individual residual signal generated for each 2 ms TTI user, adding its reconstructed signal to common residual signal.
interference from 2 ms TTI users canceled from other 2 ms TTI users’ signals
FDE: Frequency Domain EqualizerPIC: Parallel Interference CancellationRSR: Residual Stream Reconstruction
• Achievable interference reduction factor β highly dependent on:– Quality of signal that should be cancelled (2ms TTI UEs)– Data rate of UE to be cancelled– Radio channel of the UE: Multi-path profile, UE Velocity
PIC pool: •set of cells within 1 BTS that are candidates for interference cancellation (IC)•supports up to 6 cells•3 cells may perform IC simultaneously•PIC pool configuration done by operator via BTS configuration•max. 4 PIC pools per BTS•AssignedPICPool indicates which PIC pool the cell belongs to •Basic PIC functionality takes fixed number of CE per PIC pool: 48 CE•PIC-state of a cell in a PIC-Pool can be changed by AdminPICState*.
• “PIC-deactivated”, “PIC-activated”, “PIC-automatic”• PIC state change of cells with “PIC-Automatic” is
controlled by BTS• Cells with highest traffic shall be selected for IC• Cell are deselected for IC if traffic has decreased
f1 f2
cells in PIC pool
cells performing InterferenceCancellation
f1f1
f2 f2
AdminPICStateWCEL; 0 (ActivateEnabled), 1
(DeactivateDisabled), 2 (Automatic)
AssignedPICPoolWCEL; 0 (off); 1; 2; 3; 4
*There may be restriction in WBTS for changing the PICState. If the change is not possible, then the PICState remains.
Brief description of 16 QAM in UL Dual 4PAM modulation is used (4PAM→ 4 symbols & 2 bits per symbol) Variable SF≥2 for Bit Rate (BR) adjustment Multicode operation is needed to maximise Bit Rate after max SF is used
(max 4 data codes in parallel when no DPDCH configured) With W=3.84Mcps→Symbol rate=2·W/SF2+ 2·W/SF4=5760 ksps BRmax=2·5760ksps=11520 kbps
Motivation & benefits Using higher order modulation, more symbols can be transmitted, therefore
more bits can be assigned to each symbol, while the duration of symbol is kept. This results in higher Bit Rate.
Most beneficial with low intra-cell interferences. Interference cancellation techniques are welcome to lower the intra-cell
noise.
Drawbacks Reduction of the Euclidean distance between adjacent symbols. This
results in stricter requirement in SNR per symbol to achieve the same BER RoT limit requirement rises high, as the own signal interferences from
16QAM are high. Therefore this feature is dedicated for micro, pico cells.
Control data = 3.72% of whole transport block• When the transmission error occurs one small RLC PDU
needs to be retransmitted
19 MAC-es/e headers required + optional padding
19 fixed RLC PDUs (656 bits each) required per 1500 bytes IP packet
One MAC-is/i header required + optional padding
One RLC PDU is required per 1500 bytes IP packet
Control data = 0.27% of whole transport block• It corresponds to 93% drop of control data for typical IP
packet size of 1500 bytes• When the transmission errors occur one big RLC PDU
needs to be retransmitted
• Included in RU40 basic software package – no license needed• HW prerequisites: Flexi rel.2• Can be used if: Flexible RLC in DL and Basic HSUPA are both active
HSUPA Downlink Physical Channel Power Control: RAN971
• Included in RU40 basic software package – no license required• HW prerequisites: Flexi rel.2• Can be used if: Basic HSUPA, HSUPA BTS Packet Scheduler and HSUPA Basic RRM
are active
Brief description:• This feature introduces power control for following
downlink physical control channels:– E-DCH Absolute Grant Channel (E-AGCH)– E-DCH Relative Grant Channel (E-RGCH)– E-DCH Hybrid ARQ Indicator Channel (E-HICH) – Fractional Dedicated Physical Channel (F-DPCH)
• Controlling the transmit powers of the HSUPA downlink control channels based on the feedback received from UE
Motivation and benefits:• Reduction of average downlink power need • The coverage area for 2ms E-DCH TTI may be increased• Increased number of CS Voice over HSPA (RU20 – 72 users, RU40 – 128 users)
HSUPA Downlink Physical Channel Power Control: RAN971
Transmitted Power of HSUPA Downlink Physical channels is controlled through Power Offsets between HSUPA DL channels and:
•CPICH – RU10•DL DPCCH – RU20 & RU30
NodeB Tx Power
0.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
Distance from NodeB [km]
Tx
po
wer
[d
Bm
]
CPICH E-AGCH E-EGCH/E-HICH
E-AGCH Power Offset
E-RGCH Power Offset / E-HICH Power Offset
RU10
Transmitted Power of HSUPA Downlink Physical channels is controlled dynamically.Value of Tx power is changed and is incremented for the UE which are at a bigger distance from the NodeB
Without the Dynamic HSUPA BLER feature the BLER target evaluation is the same regardless to:• UE - BTS distance (cell edge / close to the
BTS)• Bursty / continuous data transmission
With the Dynamic HSUPA BLER feature the BLER target is optimized for different user scenarios and radio conditions:• Close to BTS: optimizing BLER to get peak
rates • Cell edge continuous data transmission:
optimizing radio coverage and cell capacity• Bursty traffic: optimizing latencyOLPC algorithm (RNC) enhancement Support for different BLER targets adapted to current radio transmission conditions
UL Gating (UL DTX): reduces UL control channel (DPCCH) overhead • no data to sent on E-DPDCH or HS-DPCCH UE switchs off UL DPCCH• DPCCH Gating is precondition for other 3 sub-features
CQI Reporting reduction:• CQI Reporting Reduction reduce the Tx power of the UE by reducing the CQI reporting; this
means to reduce the interference from HS-DPCCH in UL when no data is transmitted on HS-PDSCH in DL
• Reduced CQI reporting takes place only if the CQI reporting pattern defined by the last HS-DSCH transmission and CQI cycle overlaps the UL DPCCH burst of the UE DTX pattern
• N2msCQIDTXTimer: defines the number of subframes after an HS-DSCH reception, during which the CQI reports have higher priority than the DTX pattern. RNCRNHSPA; 0 (0), 1 (1), 2 (2), 4 (3), 8 (4), 16 (5), 32 (6), 64 (7), 128 (8), 256 (9), 512 (10), Infinity (11); 64 (7) subframes
• N2msCQIFeedbackCPC: defines the CQI feedback cycle for HSDPA when the CQI reporting is not reduced because of DTX. RNCRNHSPA; 0 (0), 2 (1), 4 (2), 8 (3), 10 (4), 20 (5), 40 (6), 80 (7), 160 (8); default: 10 (4) 8 (3) ms; Note: Bigger CQI reporting cycles 10ms are not recommended.
• Reduced CQI reporting takes place only if the CQI reporting pattern defined by the last HS-DSCH transmission and CQI cycle overlaps the UL DPCCH burst of the UE DTX pattern
CQI Reporting Reduction reduces the CQI reporting when there are no data transmitted on HS-DSCH for a longer period of time
ACK/NACK transmission
CQI transmission
CQI period 2ms
CQI period 4ms
CQI period 8ms
CQI transmission time defined by CQI period, but not overlapping with DPCCH transmission no CQI transmission
CQI Transmission
DPCCH pattern
UE_DTX_cycle_1 UE_DTX_cycle_1
UE_DTX_cycle_2 UE_DTX_cycle_2
7.5slots
HS-DSCH reception CQI_DTX_TIMER
UE_DTX_cycle_2
CQI_DTX_Priority set to 1
CQI_DTX_Priority set to 0
N2msCQIFeedbackCPCCQI feedback cycle (when CQI reporting not reduced)RNCRNHSPA; 0, 2, 4, 8, 10, 20, 40, 80, 160 ; 108 ms
During E-DCH inactivity, E-DPCCH detection happens at the BTS only every MAC_DTX_Cycle subframes. It is stopped at Node B after MAC_inactivity_threshold subframes of E-DCH inactivity. As a consequence, the UE experiences a delay regarding the transmission start time. The UE-specific offset parameter UE_DTX_DRX_Offset allows to stagger the processing of several UEs in time to save the BTS resources.
Discontinuous UL Reception (MAC DTX):
• N2msMACDTXCycle: length of MAC DTX Cycle in subframes. This is a pattern of time instances where the start of the UL E-DCH transmission after inactivity is allowed. RNCRNSHPA; Range: 1 (0), 4 (1), 5 (2), 8 (3), 10 (4), 16 (5), 20 (6); default: 8 (3) subframes
• N2msMACInacThr: E-DCH inactivity time in TTIs after which the UE can start E-DCH transmission only at given times. RNCRNHSPA; iInfinity (0), 1 (1), 2 (2), 4 (3), 8 (4), 16 (5), 32 (6), 64 (7), 128 (8), 256 (9), 512 (10) TTIs; default: Infinity (0)
Discontinuous DL Reception (DL DRX):
• N2msInacThrUEDRXCycle: number of subframes after an HS-SCCH reception or after the first slot of an HS-PDSCH reception, during which the UE is required to monitor the HS-SCCHs in the UE's HS-SCCH set continuously. RNCRNHSPA; Range: 0 (0), 1 (1), 2 (2), 4 (3), 8 (4), 16 (5), 32 (6), 64 (7), 128 (8), 256 (9), 512 (10); default: 64 (7) subframes
• N2msUEDRXCycle: HS-SCCH reception pattern (UE DRX Cycle) length in subframes. This parameter is a multiple or a divisor of the parameter UE DTX Cycle 1. If the value is not allowed, the parameter value minus 1 is used to calculate a new value, and so on. RNCRNHSPA; Range: 0.5 (0), 1 (1), 2 (2), 3 (3), 4 (4); default: 2 (2) subframes
• N2msInacThrUEDRXCycle: number of subframes after an HS-SCCH reception or after the 1st slot of an HS-PDSCH reception, during which the UE is required to monitor the HS-SCCHs in the UE's HS-SCCH set continuously; UE DRX Inactivity threshold; RNCRNHSPA; 0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512; 64 subframes• N2msInacThrUEDRXCycle: HS-SCCH reception pattern (UE DRX Cycle) length in subframes; RNCRNHSPA; 0.5, 1, 2, 3, 4; 2 subframes
• New parameter introduced to control step size for DL Inner Loop PC
Power Control
CPC & Power Control
TPC command
DownlinkInnerLoop PCStepSize
RNCRNAC : 0.5..2; 0.5; 1 dB
DLInLoopPCStepSizeCPCRNCRNSPA: 0.5..2; 0.5; 1.5 dB
DLInLoopPCStepSizeCPC: used by the WCDMA BTS to calculate the power increase/decrease step size when receiving TPC commands. It is applied when CPC (UE DTX, etc.) is activated for the UE.
Note: If CPC is not used for a UE, BTS applies DownlinkInnerLoopPCStepSize
• Load based AMR selection algorithm not used while CS Voice is mapped on HSPA
Requirements
CS Voice Over HSPA (RAN1689)
BG: BackgroundIA: Interactive
Codecs supported for CS Voice Over HSPA:• AMR (12.2, 7.95, 5.9, 4.75), (5.9, 4.75) & (12.2)• AMR-WB (12.65, 8.85, 6.6)
HS-DSCH
E-DCH
for Voice, SRB & other services
HSPAQoSEnabledRNCWCEL; 0..4*; 1; 0 = disabled0 = QoS prioritization is not in use for HS transport1 = QoS prioritization is used for HS NRT channels2 = HSPA streaming is in use3 = HSPA CS voice is in use4 = HSPA streaming & CS voice are in use
* if HSPA streaming or CS voice is activated, then QoS prioritization for NRT HSPA connections is in use, too
QoSPriorityMappingRNCRNPS; 0..15; 1; 14 for CS Voice over HSPA• Priority must be lower than SRB (15)• Priority must be higher than Streaming 13)
• CS voice over HSPA license exists & state is 'On‘
• HSDPA with Simultaneous AMR Voice Call license exists & state is 'On'• HSUPA with Simultaneous AMR Voice Call license exists & state is 'On'• AMRWithHSDSCH & AMRWithEDCH: HSPA with Simultaneous AMR Voice Call
enabled
• HSDPAenabled & HSUPAenabled : HSPA enabled in all Active Set cells
• HSDPA Dynamic Resource Allocation license exists & state is 'On‘• HSDPADynamicResourceAllocation is enabled
• QoS Aware HSPA Scheduling license exists & state is 'On‘• HSPAQoSEnabled is set to “HSPA CS voice” in all Active Set cells
• CPC & Fractional DPCH licenses exists & state is 'On‘• CPCEnabled in all Active Set cells• FDPCHEnabled: Fractional DPCH enabled in all Active Set cells
Enabling the feature: CS Voice Over HSPAPre-conditions
• Two different voice transmission scenarios are being considered with IP:– VoIP – UE connects with network as in standard Packed Data transmission and by using “web
communicators” a connection can be established (hard to establish appropriate charging schemes)
– CS voice over IP – voice is being carried by HSPA transport channels transparent for the user
[REF. WCDMA for UMTS – HSPA Evolution and LTE, HH AT]
• In UL there is a so called Dejitter buffer implemented in RNC PDCP• used to align the UL data stream before routing to MSC or MSS system
• In DL MAC-ehs is used to support flexible RLC PDU sizes• supporting different AMR rates
DCH
CS Core
TM RLC
RAN
CS Voice over DCH
Dejitter buffer
UM RLC
PDCP
HSPA
CS CoreRAN
CS Voice over HSPA
• Inter system mobility between 2G & 3G is as today, the CS Voice Over HSPA is just RAN internal mapping and it is not visible outside of the RAN. Handover signaling is not affected and RAN provides the measurement periods for UE using compressed mode as today
• AMR rate adaptation can be used to provide even higher capacity gains by lowering the AMR coding rate
• Voice related RRM algorithms like pre-emption are expanded to cover also the Voice Over HSPA• Air interface capacity gain of the feature depends on parameterisation of HSUPA including CPC
PtxTargetTot is calculated whenever a NCT connection is admitted
NCT: Non-Controllable Traffic
Dynamic target power is used when in cell there are SRBs or conversational services (NCT load) mapped to HS-DSCH transport channel. Dynamic target power varies between PtxTargetTotMin & PtxTargetTotMax depending on the mix of services mapped to DCH & HS-DSCH transport channels.
However, NCT load caused by services mapped to DCH transport channels must still stay below PtxTarget.
Power margin between PtxCellMax & PtxTargetTotMax is needed to protect the already admitted services mapped to HS-transport channels by giving time for the overload control to adjust PS DCH load before high priority HS-DSCH load is affected.
HSUPA Non-Scheduled Transmission NST• NST is used for the UL of CS Voice over HSPA• HSUPA TTI = 2 ms 1 HARQ process is allocated for the E-DCH MAC-d flow• EDCHMuxVoiceTTI2 & EDCHMuxVoiceTTI10 define whether or not other E-DCH MAC-d flow data
can be multiplexed within the same MAC-e PDU as CS Voice• The max. Number of Bits per MAC-e PDU for NST indicates the number of bits allowed to be
included in a MAC-e PDU per E-DCH MAC-d flow configured for non-scheduled transmissions• Generally the MAC-d flow of the SRB has higher SPI value, being prioritized over the CS voice in
the E-TFC selection• The max. SRB bit rate will be limited so that the at least 1 CS voice frame can always transmitted
together with the signaling when the max. puncturing is applied, for minimizing the CS voice delay• 2 ms TTI is selected whenever possible, otherwise 10 ms TTI is used
The maximum target value for the RTWP in UL for CS speech service allocation: PrxTargetMax defines the max. target value for the RTWP in the UL resource allocation for the CS speech services. A dynamic target of RTWP is applied in the resource allocation for the CS speech services and for the establishment of the link. Dynamic target is the closer to the value of this parameter, the less there is PS NRT R99 data traffic and RT data R99 and HSPA traffic in the cell. Establishment of the stand alone signaling link or a single service CS speech can be admitted in UL even the received non-controllable interference exceeds the value of the parameter "Target for received power" so long as the RTWP keeps below the dynamic target value defined with this parameter. WCEL: 0..30 dB; 0.1 dB; 465535 dB
Smart phones with many applications, requiring frequent transmission of small amount of data# (always-on)
To save battery power, 3GPP defines transition from states with high power consumption (Cell_DCH, Cell_FACH) to those with low consumption (Cell_PCH, URA_PCH)
approx. battery consumption in different RRC states: • Idle = 1 (relative units)• Cell_PCH < 2*1
• URA_PCH ≤ Cell_PCH*2
• Cell_FACH = 40 x Idle• Cell_DCH = 100 x Idle
*1 depends on DRX ratio with Idle & mobility*2 < in mobility scenarios, = in static scenarios# e.g. sending frequent ‘polls’ or ‘keep-alives’
3GPP Rel. 8: Fast Dormancy• modifying SCRI message; new cause value indicating
packet data session end
• RNC can keep UE in RRC connected mode, moving it into CELL_PCH/URA_PCH
UE battery life remains prolonged because power
consumption in CELL_PCH/ URA_PCH is low Network again in charge of RRC state; clarification of
“signaling connection failures” Reduction of signaling load & latency times
Cause value of ‘UE Requested
PS Data Session End’ defined
3GPP TS 25.33110.3.3.37a Signalling Connection Release Indication Cause„This IE is used to indicate to the UTRAN that there is no more PS data for a prolonged period.“
RAN2136: Fast Dormancy (FD) • Basic SW; no activation required; enabled by default• MSActivitySupervision to be configured with value > 0 to enable PCH states• Enabling FD results in T323 being broadcast within SIB1
T323:• Inclusion of T323 within SIB1 allows UE to detect that network supports FD• Setting a min. delay between 2 SRCI messages for FD; prevents, that UE is sending a flow of SCRI
messages, if network is temporarily unable to move UE to a battery-saving state
MSActivititySupervisionRNC; 0..1440; 1; 29 min
SRCI: Signalling Connection Release Indication
T323RNC; 0..7; 1; 10 s
(hardcoded)
Fast Dormancy - RNC Actions:After receiving SCRI message with cause value ‘UE Requested PS Data Session End’:• FD functionality overrides inactivity timers• RNC instructs UE to make state change to CELL_PCH/URA_PCHIf RNC receives an SCRI message without a cause value then the existing legacy functionality is applied & the UE is moved to RRC Idle mode
• Included in RU40 application software package – license is required
Brief description:• Identifies legacy Fast Dormancy phones which cause unnecessary signaling load• Provides with better network resources utilization due to shorter inactivity timers• Less signaling load because LFD (Legacy Fast Dormancy) Phones are being forced
to stay in Cell_PCH
Benefits:• Signaling load reduction on Iub, UU and Iu interfaces • Signaling load reduction in the RNC• Longer UE battery life
Overview:
SIB1 contains info about T323
• RAN supports Fast dormancy• Application has no more data to transfer• UE wants go to more battery efficient RRC state
SCRI
RNC: Data session endedRNC: UE move to more battery efficient state
Fast Dormancy Profiling: BackgroundLegacy Fast Dormancy phone detection:• The UE is detected as Legacy Fast Dormancy phone (LFDphone) when network receives
RRC:Signaling Connection Release Indication without any cause• If the Fast Dormancy Profiling feature is activated then RRC state transition is performed
according to Fast Dormancy functionality
Handling the PS Connection Establishment:• The LFD Phone after sending SCRI without any cause may still silently goes to Idle• After receiving RRC: Initial Direct Transfer, RNC checks if Iu-PS connection already exists• If yes, then all existing PS RAB resources locally and the old Iu connection are released• New Iu connection is established for pushing RRC: Initial Direct Transfer to SGSN
SCRI - without any cause RNC checks if the license is ON
If the license is available - Go to Cell_PCH
RRC: Initial Direct TransferRNC checks if Iu-PS connection for this UE already exists
Blind IFHO & Layering in State Change extends existing redirection scenarios. Both triggers utilize blind IFHO mechanism (w/o CM activation).
Target can be IF-neighbor in same or different BTS / frequency / frequency band.
IF-neighbour must have higher priority ( preference score) than source cell.
MBLB adds IF-RACH measurements to have target cell info before blind HO decision*.
Intra-freq. RACH measurement quantity (SIB11/11bis/12) to be modified from EcNo to RSCP.RSCP is used as pathloss equivalence indicating UE position within a cellRSCP (Target Cell) must be > BlindHORSCPThrTarget
DCH/HSPA cell (band Z)
CTS: Channel Type Switching* requires UE Rel. 6 or newer
MBLB: HSDPA Inactivity or Mobility triggered modes
DCH/HSPA cell (band X)
DCH/HSPA cell (band Y, RF1)DCH/HSPA cell (band Y, RF2)
DCH/HSPA cell (band Z)HSDPA traffic inactivity (instead CTS to FACH)
CM IF-HO to higher priority layer cell, on:
:
Extending event triggered scenarios (on HSDPA traffic Inactivity) & adding new HSDPA mobility triggers.
Target cell can be any Intra-RNC, IF-neighbour cell (same/different BTS, same/different freq. band) which is not in overload state1). It must have higher priority ( preference score) than source cell.
HSDPA Inactivity triggered when for UEs last active PS NRT MAC-d flow & corresponding UL PS NRT DCH/E-DCH MAC-d flow can be released
Mobility triggered HO: Adding new cell to AS (Event 1A/1C), with different preferred layer def. than the currently used. Removing cell from AS (Event 1B/1C) which has preferred layer def. currently used by UE SRNC relocation completed UE detected to have high mobility (using criteria for URA_PCH)
Each factor represented by weight is OAM configurable. Candidates for MBLB are IF neighbours on freq. layer higher prioritised then source cell layer, in bands supported by UE
Each combination „UE capability – Used/Requested Service” may be redirected by Operator to preferred frequency. E.g.: MIMO capable UE with NRT RAB should be redirected to frequency with MIMO configured cells. •Preferences are defined in new object PFL (Preferred Frequency Layer);•up to 8 carriers per „UE capability-Used Service” pair (PrefLayerXXX[1-8] = preferred RF freq.
Operator can promote one frequency band among used by neighbours of source cell. The Freq Band weight of candidate cell carrier is determined from its absolute RF number. Candidate cells in promoted band are assigned with non-zero weight: LaySelWeightBand
Freq Band weight
Band ID Frequency band (DL range)
1 RF band I (2110÷2170MHz) / IMT
2 RF band II (1930÷1990MHz) / PCS
3 Band III (1805÷1880MHz) / DCS
4 Band IV (2110÷2155MHz) / AWS
5 Band V (869÷894MHz) / CLR
6 Band VI (875÷885MHz)
7 Band VII (2620÷2690MHz) / IMT-E
8 Band VIII (925÷960MHz) / GSM
.. …
11 Band XI (1475.9÷1495.9MHz)
…
LaySelWeightBandPFL; 0..10000; 1; 0 (not used)
PreferBandForLayeringPreferred UMTS band
RNMOBI; 0..51; 1; 0 (no band pref.)
if any band preferred PreferBandForLayering
Freq.BandWeight (of pref. band freq.) = LaySelWeightBand
Operator can prioritise IFHO to lower/higher band in case RSCP is below/above defined thresholds. RSCP of source cell* is used as indication of UE distance from source cell. Close UE can be moved to higher band (smaller cell range), while distant UE to lower band to improve radio link situation
RSCPweight
* Source cell: best RSCP cell in RACH Intra-freq report
Low Band ( F1)
High Band (F2)
21
if RSCP (Source Cell) ≥ BlindHORSCPThrAbove
RSCPWeight (higher band freq.s) = LaySelWEightRSCP
• These parameters can be used to optimize frequency usage in case of co-located multi-band cells: – UEs close to BTS allocated to a high band – Low band capacity used for distant UEs RSCPWeight applied only to following events:
• Blind HO in RAB Setup phase• State transition to Cell_DCH state
The Load weight is used to avoid IFHO to more loaded cells or direct traffic to balance the load. The weight is decided based on mean HSDPA power per HSDPA NRT User.
UE detected to be fast moving UE with same mechanism than for URA-PCH (too many mobility events per time)• In CELL_DCH state, UE location is known in cell level and handover process can
calculate the velocity of the UE by AS changes needed for the UE. • UE is fast moving: if number of complete AS changes ≥ FastUEThreshold during
time period FastUEPeriod
• UE fast moving preference score is calculated taking into account
only the preferred layer definition for fast moving UE:• PreflayerFastMovUEPS is used for PS, and• PreflayerFastMovUECS is used for CS
• fast moving UE can have only 2 preferred frequencies in priority order
FastUEPeriodWCELRNMOBI; 01..60; 1; 10 s
FastUEThresholdWCELRNMOBI; 2..60; 1; 3
AS: Active Set
PreflayerFastMovUEPS#2
PFL; 0..16383; 1; -0
PreflayerFastMovUECS #1
PFL; 0..16383; 1; -0
#1 defines the preferred layers for fast moving UE with AMR RAB or other CS RAB, 0...1 streaming RAB + 0...3 NRT RAB(s) allocated#2 defines the preferred layers for fast moving UE with 0...3 PS NRT RAB(s) + 0...1 PS streaming RAB allocated. (No AMR RAB or other CS RABs exists)
Multi-Band Load Balancing MBLBMBLB / PFL Example – RSCP Weight
RSCPWeight is applied only to the following events:• Blind HO in RAB Setup phase• State transition to Cell_DCH state
1. UE is moved to higher band when path loss of serving cell is low • RSCP(S_cell) > BlindHOThrAbove (e.g. -72 dBm). Higher frequency band, which get non-zero
RSCP weight value (LaySelWeightRSCP ≠ 0), are preferred.
2. UE is moved to lower band when pathloss of serving cell is high • RSCP(S_cell) < BlindHOThrBelow (e.g. -92 dBm). Lower frequency band, which get non-zero
High Speed Cell_FACH (DL): RAN1637• Included in RU30 application software package – license required• HW prerequisites: Flexi rel.2• Can be used if: Flexible RLC Downlink is active
Brief Description: • This feature enables Fast Cell_PCH to Cell_FACH switching (transition <200ms)• Feature reduces signaling load on Iub and Iur interfaces• Reduces code tree occupation• Saves BTS baseband resources• Increases number of supported smartphones• Increases possible throughputs on common channels to 1.80Mbps in DL
• Included in RU40 application software package – license required• HW prerequisites: Flexi rel.2• Can be used if both Flexible RLC Downlink and Flexible RLC in Uplink features are active
Brief Description: • This feature enables Fast Cell_PCH to Cell_FACH switching (transition <100ms)• Feature enhances High Speed Cell_FACH in DL• Increases possible throughputs on common channels to 1.45Mbps in UL
RU20 uses HSDPA BLER Targets of 10 % & 25 %:– 10 % is applied in static channel conditions– 25 % is applied in fading channel conditions– these BLER targets are not configurable & independent CQI (high or low)
RU30 introduces Dynamic HSDPA BLER•allows the use of HSDPA BLER Targets of 2 %, 6 %, 10 % & 25 %•the Ru30 BLER Target is a function of the:
the channel conditions (static vs fading) the CQI
•Thresholds defining high, medium & low CQI ranges are configurable•Upper & lower BLER target limits for each CQI range are configurable
•Dynamic HSDPA BLER is a basic software feature; it requires no license•The feature is installed as part of the BTS software and is always enabled•Fallback to RU20 behavior: configuring BLER Target parameter set appropriately•Parameters associated with this feature are Node B commissioning parameters rather than RNC databuild parameters
•Motivation/Benefits: cell & end-user HSDPA throughputs improved by up to 8 %
Dynamic HSDPA BLER: Upper & lower BLER Target limitsThe table below presents the RNC databuild parameters used to define:• the set of 3 CQI ranges• the upper & lower BLER Target limits for each CQI range
Condition for each CQI range Upper & Lower BLER Target Limits*